| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Molecular Biology, University of Wyoming, Laramie, Wyoming, USA
1Correspondence: Department of Molecular Biology, University of Wyoming, Laramie, WY 82071, USA. E-mail: jordanka{at}uwyo.edu
 |
ABSTRACT |
|---|
Key Words: histone chaperone • nucleosome assembly • transcription
|
GENERAL INTRODUCTION TO NAP1 |
|---|
. This histone chaperone/nucleosome assembly property of Nap1 has been used extensively in the chromatin field to reconstitute nucleosomal templates for a variety of structural and functional studies.
Nap1 is highly conserved among eukaryotes. Nap1 homologues are present in yeast (2)
, Drosophila (3)
, Plasmodium (4
, 5)
, nematodes (6)
, soybean (7)
, tobacco and rice (8
, 9)
, Xenopus (10
, 11)
, mouse, and human (1)
. The yeast NAP1 gene is nonessential (12)
, but deletion of the NAP1 gene in Drosophila and mouse leads to embryonic lethality (13
, 14)
. Nap1 is a representative of a large protein family whose members show limited sequence homology (for further references, see ref. 11
). Some members of the family exhibit tissue-specific expression patterns (e.g., refs. 11
, 14
). Nap1 has been implicated in a large variety of seemingly unrelated cellular functions, with a vast number of protein partners identified by both traditional and proteome-wide approaches. To give the reader an idea of the complexity of possible functions of Nap1, we have listed these partners in Table 1
. Whether the described interactions are physiologically relevant remains to be seen: it is clear, though, that understanding the mechanisms behind these interactions and their precise role will require a lot of experimental effort.
|
We chose to focus this review on the numerous chromatin-related functions of Nap1 and how the protein may participate in the process of transcription, both on naked DNA and on chromatinized templates. This focus requires a brief introduction to chromatin structure and the definition of some terms that will be used throughout the review.
|
THE NUCLEOSOME AND CHROMATIN: DEFINING TERMINOLOGY |
|---|
consisting of a nucleosome core particle and linker DNA. The core particle itself contains 147 bp of DNA wrapped in a left-handed (negative) sense around a histone octamer consisting of two molecules each of histones H2A, H2B, H3, and H4 (44
45
46
47)
(Fig. 1
). The octamer is a modular structure possessing 2-fold symmetry: the H3/H4 tetramer is at the center, with the two H2A/H2B dimers attached on either side. The tetramer organizes the central portion of nucleosomal DNA, while each H2A/H2B dimer contacts 
30 bp on either side. We will often refer to particles containing a complete octamer as octasomes without any reference to the length of DNA (
147 bp). The H3/H4 tetramer is capable of wrapping the DNA around itself, and the sense of DNA wrapping could be either negative (as in the complete octasome) or positive (49)
. This particle will be termed a tetrasome, without any reference to the particular sense of DNA wrapping. The H2A/H2B dimers cannot form a particle by themselves, although they do bind to DNA. If only one H2A/H2B dimer is attached to the tetramer in the particle, we will talk about a hexasome. Figure 1
schematically presents these particles.
|
We will be using the term nucleosomal array to define a "beads-on-a-string" type of arrays of successive nucleosomes not containing linker histones. Linker histones are a small family of isoforms; they bind to the linker DNA entering and exiting the core particle of the nucleosome, forming the so-called chromatosome (Fig. 1)
. Nucleosomal arrays can be obtained from isolated chromatin fibers by removal of linker histones or, alternatively, by reconstituting core histones onto long linear or circular DNA molecules. In contrast, the term chromatin fiber will only be used to denote linker histone-containing nucleosomal arrays (whether isolated from nuclei or obtained in vitro by reconstituting linker histones to nucleosomal arrays).
|
NAP1 STRUCTURE, POSTSYNTHETIC MODIFICATIONS, INTRACELLULAR LOCALIZATION, AND OLIGOMERIZATION STATES |
|---|
. The protein is highly acidic, with 25% of amino acid residues being negatively charged at physiological pH. These residues are clustered in four regions, two of which are preceded by highly evolutionarily conserved regions of unknown function. The protein contains both a nuclear localization signal and a nuclear export signal. The recently published 3 Å crystal structure (50)
identifies a unique 3-dimensional fold composed of 
helices and ß strands (Fig. 2)
. There is a single disulfide bond between Cys-249 and Cys-272 that covalently links the ß3 and ß4 strands. The long 2 helix is responsible for homodimerization by close antiparallel pairing of the 2 helix from each monomer along their entire length. This is an unusual arrangement that differs from the common "coiled-coil" motif found in many proteins. The dimer has an overall ellipsoidal shape, with an uneven charge distribution: the concave underside is highly acidic, but the cavity formed is too small to accommodate even an helix. Park and Luger (50)
suggest that this acidic surface may be involved in neutralizing and binding the basic amino-terminal histone tails (see below). The antiparallel ß sheet formed by ß1–ß4 strands seems to be a conserved feature in many of the histone chaperones whose structure has been determined or predicted (nucleoplasmin, dCAF1 p55, Asf1).
|
Nap1, as well as its close relative Nap2, are phosphoproteins that undergo cell cycle-dependent phosphorylation-dephosphorylation events. The kinase involved in phosphorylation has been identified as casein kinase II (CKII), and direct interactions of CKII and Nap1/2 have been reported (6
, 41
, 42)
. Yeast and Xenopus Nap1 can be phosphorylated in vitro by cyclin B/p34cdc2 kinase complexes (note that only B cyclins, but not A cyclins, interact with Nap1; see Table 1
) (10)
. Nap1 is acetylated by p300, probably in a cell cycle-dependent manner, since the formation of the Nap1-p300 complex increases during S-phase (51)
. Finally, the protein is subject to an unusual postsynthetic modification: polyglutamylation (52)
. This modification, originally identified on tubulin, consists of side chains of several glutamyl residues added onto two putative glutamylation sites, Glu-356 and Glu-357, in the carboxyl-terminal domain of Nap1. The functional significance of these postsynthetic modifications remains to be determined. They may have a role in determining the equilibrium between different Nap1 oligomeric states and in Nap1 transport between the cytoplasm and the nucleus (see below).
The intracellular localization of Nap1 has been debatable. Studies of Drosophila embryos indicated nuclear localization during S-phase and predominantly cytoplasmic localization during G2 (53)
. In general, Nap1 and Nap2 are mainly cytoplasmic in G1 and G2, and translocate into the nucleus during S-phase (21
, 51
, 54)
. Of the three electrophoretic bands that can be detected in HeLa cytoplasm with Nap1 antibody, only the upper band is prominent in the nucleus (55)
; these bands may represent differently modified (phosphorylated?) forms of Nap1. Asahara et al. (51)
, however, found only two anti-Nap1-detectable bands in HeLa, represented in equal proportions in whole-cell and nuclear extracts. Further experiments are needed to determine the exact modification state of the nuclear and cytoplasmic forms of the protein, and whether a change in modification is a condition for its nucleo-cytoplasmic transport and for its presumptive role as a chromatin assembly factor.
The exact mechanism by which Nap1 is transported into the nucleus and back to the cytoplasm is unclear. Both NLS and NES have been identified (see Fig. 2
). Genetic studies in yeast have indicated a somewhat decreased nuclear accumulation in strains lacking the gene for karyopherin 114 (the major karyopherin involved in H2A/H2B transport; see below) (38)
. The lack of complete inhibition of Nap1 import in kap114
yeast demonstrates that Nap1 has more than one route to the nucleus, although these routes remain to be identified. The back export to the cytoplasm remains uncharacterized (for more discussion, see ref. 38
).
Data scattered throughout the earlier Nap1 papers suggested that Nap1 may exist in more than one oligomeric complex. More recently, several laboratories have focused on the identification and functional characterization of Nap1’s associated states and on the stoichiometry of their complexes with histones. A complex mixture of Nap1 species under physiological conditions has been described (56)
, with the dimer further self-associating to form higher order oligomers. S. Uchiyama (personal communication) detected even-numbered oligomers all the way from dimers to octamers using MS and gel filtration chromatography after chemical cross-linking of hNap1. Toth et al. (57)
identified the prevalent species as a dimer and an octamer. They calculated the concentrations of Nap1 in the nucleus and cytoplasm, and concluded that Nap1 octamer formation could be induced in vivo in S-phase nuclei (Fig. 3
). Furthermore, they suggested that phosphorylation, acetylation, and/or polyglutamylation might affect the dimer/octamer equilibrium. As far as the stoichiometry of Nap1 binding to histones is concerned, the prevailing opinion is that each Nap1 dimer binds to one dimer of either H2A/H2B or H3/H4 (57
, 58
; S. Uchiyama, personal communication). In contrast, Nap1 binds linker histones at a 2:1 (Nap1:linker histone) molar ratio (59)
.
|
The question of whether the histones bound to Nap1 can be organized in an octameric complex has not been resolved satisfactorily. Ishimi et al. (60)
found that all four histones bound to Nap1 after a preformed H2A/H2B/Nap1 complex was incubated with free H3/H4 and suggested that, in this case, an octamer of histones formed on Nap1. A similar conclusion was reached by Rodriguez et al. (42)
regarding Nap2: eluates from Nap2-Sepharose affinity columns contained all four histones in stoichiometric amounts. Nakagawa et al. (61)
checked whether H3/H4 could be coimmunoprecipitated by anti-H2A/H2B antibodies from an equimolar mixture of H2A/H2B and H3/H4 when Nap1 was present. Only half of H3/H4 coprecipitated, in contrast to salt-reconstituted mononucleosomes, where H3/H4 were quantitatively (100%) coprecipitated. The roughly equimolar core histone ratios found in earlier work could reflect the presence of Nap1 multimers, with each Nap1 dimer binding to a histone dimer (see above) (this would require similar affinities of H2A/H2B and H3/H4 to Nap1, which have often been observed under in vitro conditions).
|
HISTONE BINDING |
|---|
, 62)
has prompted extensive histone binding studies, mostly in vitro; an important issue that has been addressed often is whether the unstructured amino-terminal tails of the histones are needed for the interaction. It is beyond the scope of this review to detail these studies, especially since many of the results were controversial. The interested reader is referred to a set of relevant publications (e.g., refs. 9
, 38
, 54
, 58
, 63
64
65
66
67
). In general, Nap1 seems to interact with all four core histones in vitro; the in vivo situation is much less clear, with only H2A/H2B or H2A being identified in complexes with Nap1 (53
, 68
, 69)
. It is relevant to mention here that Nap1 is involved in the actual transport of H2A/H2B (but not of H3 and H4) from the cytoplasm to the nucleus (70)
. In this transport-related function, Nap1 forms a complex with the major karyopherin protein involved in H2A/H2B transport, Kap114p, and serves as a bridge between Kap114p and the nuclear localization sequence in the histone amino termini. Nap1 also serves as a specificity factor, ensuring the preferential binding of H2A/H2B to Kap114p and selecting it among the numerous members of the karyopherin family (for a review, see ref. 71
).
Linker histones have long been known to bind to Nap1/2 in vitro (5
, 9
, 54
, 56
, 66)
. Recent work has used Nap1 as a linker histone chaperone to delineate functional differences between the maternal B4 linker histone variant and its somatic H1 counterpart in developing Xenopus (72)
. Nap1 was also shown to act as such a chaperone in sperm chromatin remodeling and in the context of reconstituted nucleosome dimers (73)
. Finally, Keppert et al. (59)
unambiguously demonstrated that Nap1 removes linker histones from chromatin fibers isolated from HeLa cells and induces a transition to a fiber of a more extended state. Figure 4
illustrates the chromatin structural components that bind directly to Nap1; it also lists Nap1 binding partners that may have a regulatory function in transcription through chromatin (see below).
|
|
NUCLEOSOME ASSEMBLY, OCTAMER SLIDING, AND H2A/H2B DIMER REMOVAL OR EXCHANGE |
|---|
). In vitro, Nap1 can form nucleosomes by itself without any help from additional factors. How does this happen? What is the chromatin formed by Nap1 like? Can it be that Nap1 can do this "solo" job in the cell as well under specific circumstances (e.g., on nonreplicating DNA)? While the mechanism of Nap1-mediated nucleosome formation in vitro seems fairly well defined, the in vivo relevance of such a pathway remains to be established.
Nap1 has been shown to efficiently deposit histones onto supercoiled DNA to form nucleosomes (61)
. The Nap1 domains required for nucleosome assembly are delineated in Fig. 2
alongside other identified functional portions of the molecule. The ordered transfer of the H3/H4 tetramers first, followed by the deposition of H2A/H2B dimers, is determined by differences in the relative binding affinities between the components of the system: DNA, H3/H4, H2A/H2B, and Nap1 (61)
. For H2A/H2B, the order of affinity is 1) tetrasome, 2) Nap1, 3) DNA; for H3/H4, the order of affinity is 1) DNA) and 2) Nap1. These affinities explain why H3/H4 transfer occurs first to form a tetrasome and why H2A/H2B dimers are added only to the tetrasome, not to naked DNA. Rippe’s group has investigated the process of nucleosome assembly on short mononucleosome-sized DNA fragments, 146 and 207 bp in length (75)
. Dissecting the individual assembly steps to derive their time constants, they found that a tetrasome forms first but is quickly converted to a hexasome; the addition of the second H2A/H2B dimer is the slowest step in the process, with a rate constant comparable to that of histone H1 addition. It was concluded that the specific formation of the nucleosome/chromatosome is governed simply by thermodynamic differences in the interactions among Nap1, histones, and DNA, in agreement with ref. 61
. A recent single-molecule study also indicates that Nap1-mediated nucleosome assembly is a three-step process compatible with deposition of one H3/H4 tetramer, followed by two H2A/H2B dimers (76)
.
The nucleosome assembly activity of Nap1 may be regulated in vivo. Indeed, Mosammaparast et al. (39)
demonstrate that karyopherin Kap114p, the primary importin responsible for the nuclear import of histone H2A/H2B, inhibits Nap1-mediated chromatin assembly. The mechanism involves direct binding between Kap114p, Nap1, and the histones in a complex that persists in the nucleus. Other regulators of Nap1 deposition activity are probably going to be discovered.
Structure and dynamics of chromatin assembled by Nap1
Finally, are there some specific features of chromatin fibers (or of nucleosomal arrays that do not contain linker histones) assembled entirely by Nap1? Kadonaga and Ito groups have asserted that arrays assembled by Nap1 on random DNA sequences are quite irregular, with variable distances between successive nucleosomes (e.g., refs. 3
, 61
). Nap1 efficiently forms nucleosomes at random DNA locations; the "maturation" of these irregular fibers into more regular structures with "physiological" spacing of nucleosomes occurs later through the action of ATP-dependent chromatin remodelers such as ACF (77)
.
Although the view stated above is widely accepted, some laboratories have reported that Nap1 by itself can form regular nucleosomal arrays. Laybourn’s laboratory has observed regular nucleosome array formation, with relatively short repeat length (160–165 bp), using yNap1 as assembly factor in a system containing purified yeast core histones (78
, 79)
. McQuibban et al. (66)
reported that recombinant yNap1 and chicken erythrocyte histones could produce uniformly spaced nucleosomes, with a repeat length of only 146 bp. Moreover, salt dialysis reconstitution did not produce regular nucleosome ladders on micrococcal nuclease treatment, but subsequent incubation with Nap1 created uniform spacing. The spacing activity of Nap1 occurred spontaneously at 37°C and did not require ATP hydrolysis; the histone tails were needed for the remodeling to occur as well as for the initial assembly. The observation that Nap1 could be an ATP-independent chromatin remodeling factor, at least in terms of causing sliding of the histone octamer along DNA, is of great potential importance for some nuclear processes that require factor access to specific binding sites in nucleosomal DNA.
Nap1 effects on mononucleosome particles: octamer sliding, H2A/H2B dimer removal, or exchange
This issue has been further studied in monosomal particles reconstituted on a 196 bp nucleosome positioning sequence derived from the 5S rRNA gene (80)
. Two major nucleosome species that differ with respect to their translational positions were produced on the 5S DNA fragment by salt dialysis: in one nucleosome the octamer occupied a central position; in the other it was located at the end of the fragment. Overnight yNap1 incubation, even at 4°C, transformed the entire population of centrally positioned nucleosomes into end-nucleosomes. Further experiments indicated that transient H2A/H2B dimer dissociation by yNap1 was required for sliding to occur. The human Nap1 did not show such nucleosome sliding activity when tested on the same sequence reconstituted with either canonical recombinant core histones or the H2A variant H2A.Bbd (81)
. However, hNap1 did remove H2A.Bbd/H2B dimers from mononucleosomes, indicating that dimer removal could occur without subsequent sliding (81)
. Actually, the authors used the Nap1-mediated formation of hexasomes to assay the comparative stability of nucleosomal particles containing all known histone H2A variants: no appreciable removal of dimers was observed, with the notable exception of H2A.Bbd. Another set of experiments led to the conclusion that Nap1 preferentially mediates the exchange of unstable dimers in the nucleosome for stable dimers. The recent finding from Luger’s laboratory that Nap1 can exchange "regular" H2A/H2B dimers with dimers containing the histone variant H2A.Z (80)
may reflect the higher stability of H2A.Z containing nucleosomes (82
, 83)
.
In principle, such an "exchange" capability could be of great importance for the functioning of chromatin, since the histone core variants may impart special structural and dynamic features to chromatin containing them (for recent reviews, see refs. 84
85
86
87
). Since Nap1 cannot efficiently remove canonical H2A/H2B (which is the prevalent dimer) (81)
, other factors such as ATP-dependent chromatin remodeling machineries could be involved. Recent in vitro work from Kornberg’s laboratory shows that this indeed may be the case (88)
. These authors demonstrate that the chromatin remodeler RSC can either remove both H2A/H2B dimers, with the production of a tetrasome, or totally disassemble the nucleosome, depending on which histone chaperone is present: Asf1 or Nap1. Moreover, chromatin remodelers have been implicated in histone exchange in vivo (e.g., refs. 89
90
91
92
). The ATP-dependent chromatin remodeling complex, SWR1, has been identified as histone H2A.Z exchanger both in vitro and in vivo (Swr1p is a member of the Swi2/Snf2 family of chromatin remodeling ATPases) (89
, 90
, 92)
. To confirm the physical association between H2A.Z and the SWR1 complex, Mizuguchi et al. (92)
performed immunoprecipitation of FLAG-tagged native H2A.Z from yeast whole-cell extracts. Nap1 was among the bound proteins, although its precise role in the exchange reaction remains unclear, especially since in vitro the reaction proceeded efficiently whether or not Nap1 was present.
It must be noted that other recent work (59
; C. Seebart, M. Tomschik, and J. Zlatanova, unpublished results) failed to observe such Nap1-mediated H2A/H2B dimer dissociation from native or reconstituted nucleosomal arrays despite the clear production of hexasomes in reconstituted core particles (59)
or particles containing 208 bp of 5S rDNA (M. Tomschik, C. Seebart, and J. Zlatanova, unpublished results). Thus, whether Nap1 can partially (or fully) disassemble nucleosomes in the physiologically relevant context of nucleosomal arrays/chromatin fibers remains to be seen. It may well be that in vivo Nap1 cooperates with other factors to extract dimers (see also discussion on transcription).
Figure 5
A illustrates the structural changes that Nap1 may inflict on a nucleosomal particle. Nap1 may also directly affect the DNA wrapping around the histone core without removing histone proteins, as suggested by DNase I accessibility results (81)
(not illustrated in figure). Figure 5B
depicts our current hypothesis about how Nap1 can govern the structural transitions in the nucleosome by serving as a chaperone for proteins that bind at the entry-exit point of the nucleosome. Recent studies have made it clear that the DNA in the nucleosome particle may undergo spontaneous partial unwrapping from the histone core to expose sites hidden in the particle (for a recent review of this work and a model of how chromatin DNA is scanned by protein factors, see ref. 93
) (see also Fig. 6
). This spontaneous "breathing" (short-range unwrapping) or "opening" (long-range unwrapping) of nucleosomal DNA from the ends of the particle may be obstructed by the binding of linker proteins. In the hypothetical scenario presented, the linker histone would "close" the nucleosome gate, and proteins such as HMGB1 would "keep it ajar"; other proteins such as histone H5 may "lock" the door by binding very strongly (H5 is the linker histone variant characteristically present in nucleated chicken erythrocytes, cells that are terminally differentiated with no transcription or replication). Nap1 may, in a regulated way, govern these types of transitions by exchanging one linker protein for another. This hypothesis would need to be experimentally verified.
|
|
|
NAP1 AND TRANSCRIPTION |
|---|
yeast cells indicated that 10% of all yeast open reading frames are affected by the absence of Nap1, exhibiting either increasing or decreasing expression levels (22)
. Moreover, the affected genes are found in clusters along chromosomes; these observations were interpreted in terms of a requirement for Nap1 for nucleosome maintenance. In vivo studies using transcription reporter constructs indicate that Nap1 (or Nap2) alone can augment transcription by a factor of 2 (67
, 94)
. These stimulatory effects of Nap1 are considerably elevated in the presence of activator proteins (e.g., TF E2) or coactivators (e.g., p300) (67
, 94
; see also below). In vitro transcription performed on reconstituted nucleosomal arrays using partially purified general transcription factors (TFs) and RNAP II indicated a similar stimulatory effect of Nap1 on TF-dependent transcription (51)
.
Nap1 affects transcription initiation by interaction with transcription factors
Nap1 interacts with several RNAP II general TFs as well as with sequence-specific transcriptional activators and repressors (Table 2
). The majority of these interactions have been identified using yeast two-hybrid screens, genetic, or proteomic approaches such as tandem affinity purification combined with mass spectrometry. Thus, most of these interactions need to be further investigated in order to identify false positives and to achieve an understanding of physiological relevance and underlying mechanisms. Recently, Nap1 was identified as a strong transcriptional activator in a proteome-wide search for yeast proteins with transcriptional activity (97)
. The proteins were fused carboxyl-terminally to a Gal4-DNA binding domain and expressed in a yeast strain with Gal4 binding sites in a His3 reporter gene. Although the study did not determine the actual activation domains, the fact that Nap1 was among the 132 strong activators (of 6000 investigated) certainly deserves close attention.
|
Nap affects transcription initiation at nucleosomes
Nap1 has also been implicated in transcription through chromatinized templates, both in vivo and in vitro. It has been known for years that nucleosomes present a formidable barrier to transcription at the levels of both initiation and elongation. Despite years of research, the issue of how gene-specific and general TFs and the polymerases themselves gain access to the underlying DNA template remains an enigma. It is fairly obvious that nucleosomes have to undergo some kind of structural transitions (up to complete removal) in order to allow events like promoter melting or transcription bubble movement along the DNA to take place (for further discussion and recent literature surveys, see refs. 98
, 99
). We will present evidence separately pointing to Nap1 participation in transcription initiation and elongation through nucleosomes (Fig. 6
and Fig. 7
).
|
Among the first indications that Nap1 may facilitate TF access to DNA sites hidden in the nucleosome was the observation that Nap1 stimulates binding of a fusion DNA binding protein (Gal4-AH) to reconstituted core particles, and this stimulation was dependent on octamer disruption and histone displacement (100)
(Fig. 6A
). In this scenario, TF binding to a nucleosomal site is mediated via removal of a H2A/H2B dimer by Nap1, and thus is the first step in the activation of transcription. It is worth noting that such a removal of a dimer has been observed only in isolated nucleosome particles and may not have physiological relevance in the context of intact chromatin.
Order-of-addition experiments have suggested another possible succession of events in transcription initiation on chromatin templates: 1) TFs and ATP-dependent chromatin remodelers cause an initial remodeling event in nucleosomal arrays; 2) in a second step, the transcriptional activator recruits histone acetyltransferases (e.g., p300) to promoters; 3) histone acetylation facilitates the loss of H2A/H2B dimers to Nap1 or other histone chaperones (101)
. (As noted below, p300 may not be the enzyme to acetylate histones, but rather itself and Nap1; see ref. 51
.) The effect of histone acetylation on the histone acceptor ability of Nap1 has been substantiated in subsequent work (102)
, but the exact molecular mechanism requires further investigation. Acetylation may reduce the affinity of H2A/H2B for the H3/H4 tetramer or for the DNA ends in the nucleosome (the sites of H2A/H2B binding); alternatively, acetylation may change the affinity of H2A/H2B for Nap1. The two scenarios (100
, 101)
differ as to when exactly in the initiation process Nap1 removes H2A/H2B dimers from nucleosomes: at the very beginning to help TF binding or much later in the process, long after TF binding and chromatin remodeling has occurred.
From what is known about spontaneous conformational transitions in nucleosomal particles (93)
(see also above), it is highly probable that Nap1 may change the dynamics of these transitions, specifically affecting the spontaneous partial unwrapping of the DNA ends without removal of histones. Preliminary data from our laboratory (M. Tomschik et al., unpublished results) obtained by single-pair fluorescence resonance energy transfer indicate that this may indeed be the case. The two scenarios of how Nap1 may facilitate TF binding to nucleosomes are presented in Fig. 6A
.
Nap1 may further exert a stimulatory effect on transcription initiation by physically stabilizing the binding of transcription activators or coactivators to sites within or close to a nucleosome. Indeed, direct interactions between Nap1 and E2 (a transcriptional activator encoded by papillomaviruses), as well as Nap1 and p53 (94)
, have been reported. It is possible that Nap1 stabilizes binding of the TF to nucleosomes, creating a bridge between the factor and H2A/H2B, as illustrated in Fig. 6B
(left-hand panel). Similarly, Nap1 can help in the formation of complexes involving coactivators. A well-studied example is the coactivator histone acetyltransferase p300 (51
, 67
, 101
; Fig. 6B
, right-hand panel). Stable triple complexes containing H2A/H2B, Nap1, and p300 have been identified (51
, 67
, 101)
; ternary complexes between Nap1, p300, and E2 activator have also been reported (94)
. It remains to be established exactly how these stable complexes involving Nap1 exert their effect on transcription initiation. This question is important, since Nap1 has not been reported to stably bind to chromatin (e.g., refs. 66
, 75
, 103
). It must be noted that Nap1 potently inhibits acetylation of core histones by p300, suggesting that p300 may be acetylating either Nap1 or itself (51)
.
Nap1 affects transcription elongation through nucleosomes
What about transcription elongation through nucleosomes? Is Nap1 involved, and how? Recent synthetic genetic array analysis has identified several subunits of the RNAPII Elongator histone acetyltransferase complex as Nap1 partners (96)
. Elongator is a complex of six subunits, isolated as a component of the chromatin-associated, hyperphosphorylated (elongation-active) form of RNAP II in yeast (104)
. Elongator probably is loaded onto the transcribing polymerase during promoter clearance and travels with it along the DNA of the coding region (for further references, see ref. 96
). Proteomic studies using TAP-MS have shown that the two subunits of the FACT (facilitates chromatin transcription) complex also interact with Nap1 (17)
. FACT has been identified as a chromatin-specific elongation factor required for transcription of chromatin templates in vitro and in vivo (105)
. The human FACT comprises two protein subunits: Spt16 (the human homologue of the essential yeast protein Spt16 implicated in transcription) and Ssrp1 (a high-mobility group 1-like, structure-specific recognition protein). FACT interacts with both nucleosomes and isolated H2A/H2B dimers, and can assemble nucleosomes in vitro (106
; reviewed in refs. 107
, 108
). Thus, FACT may be involved in two distinct elongation-related functions: disassembly of nucleosomes in front of the elongating polymerase by removing an H2A/H2B dimer from the octasomes and assembly of complete nucleosomes in the wake of the enzyme (107
, 108)
. Presumably, Nap1 could perform exactly the same functions in transcription elongation, nucleosome disassembly, and assembly (see next paragraph). Its high abundance (109)
and the fact that it is a single polypeptide may actually make it the preferred protein for such functions. What the physiological relevance of the interactions between two factors with similar capabilities—Nap1 and FACT—may be remains unclear and needs further experimentation to be explained at the mechanistic level. Finally, a genetic interaction between nap1 and acf1 (the gene that encodes a component of the chromatin remodeler ACF) has been identified in Drosophila (110)
. This link between Nap1 and Acf1 is worth pursuing further.
A wealth of information about the possible involvement of Nap1 in transcription elongation comes from a series of careful in vitro transcription experiments in model systems that use polynucleosomal templates reconstituted on linear or superhelical DNA molecules of different levels of superhelical stress (102
, 111
, 112)
. Combined with previous knowledge about 1) a clear preference of H2A/H2B to bind to (–) supercoiled DNA and to left-handed tetrasomes; 2) a preference of H3/H4 to bind to (+) supercoiled DNA; and 3) preference of H3/H4 to bind to RNA over DNA (reviewed in ref. 98
), Jackson and co-workers have come up with a model of how nucleosomes are disassembled and reformed during transcription and what the role of Nap1 may be. The general idea is illustrated in Fig. 7
. An original and important point needs to be emphasized here. It may be that nucleosome disassembly in the path of the elongation polymerase takes place via two different mechanisms early or late during transcription, and this difference is determined by the very different levels of superhelical stress accumulating along the chromatin template as a result of transcription. Every 10 bp transcribed creates, in a topologically constrained DNA template, one (+) superhelical turn in front of the polymerase and one (–) superhelical turn in its wake (this twin domain model of transcription was suggested by Liu and Wang in 1987 and subsequently proven experimentally beyond a doubt; ref. 113
). This phenomenon will create a large differential in the level of superhelical stress along a gene during transcription, being especially prominent in long genes (but still large even in shorter templates). This scenario assumes that topoisomerases are unable to rapidly remove the accumulated stress, an assumption that has found recent experimental support (114)
. Thus, the role of Nap1 will be different early and late during transcription: early in transcription Nap1 will extract the first dimer; then, in cooperation with the second dimer, will help transfer the H3/H4 tetramer to the nascent RNA chain. The H2A/H2B/Nap1 complex will be available to put the H2A/H2B cargo back to reforming particles in the wake of the polymerase. At later points, disassembly of the octasomes will occur as a result of the high (+) superhelical stress (with the tetramer never leaving the template); the role of Nap1 will be to reform the nucleosome by adding H2A/H2B dimers to the tetrasome left behind the polymerase. Jackson’s model is based entirely on in vitro experiments; its physiological relevance remains to be determined.
Another potential link to supercoiling and transcription comes from the observation that Nap2 coeluted from chromatographic columns with Topo I (115)
. Topo I and Topo II enzymes have long been implicated in transcription through nucleosomes. More recent work from Parvin’s laboratory has indicated that core-RNAP II activity is inhibited on nucleosomal templates whereas that of holo-RNAP II is not markedly repressed (116)
; this difference was attributed to the presence of Topo II in the holo-enzyme. These in vitro results agree with the well-documented role of topoisomerases in eukaryotic gene transcription in vivo (e.g., ref. 117
). The Nap1/2-topoisomerase connection certainly deserves further attention.
From what has been said, the participation of Nap1 in transcription in vivo seems probable. Apart from that, our knowledge is extremely limited and substantial experimental effort is needed to gain insight into the scenarios and mechanisms of Nap1 action.
|
CONCLUDING REMARKS |
|---|
Once chromatin is assembled, Nap1 seems to be capable of at least some activities that provide "fluidity" to the nucleosome particle, nucleosomal arrays, or "complete" chromatin fibers. These may include H2A/H2B dimer removal or exchange and sliding of the octamer along the DNA (which may require dimer removal). The hints that Nap1 could be an ATP-independent chromatin remodeling factor, at least in terms of causing nucleosome sliding, should be investigated further. Can Nap1 go all the way back to totally dismantling the particle in a reversal of its assembly activity? Based on measured relative affinities, removal of H3/H4 seems unlikely, but factors present in vivo (e.g., RSC) may help in this process.
Nap1 may also perform another important function: controlling association and exchange of proteins that bind on the outside of the particle at or close to the point where DNA enters and exits the nucleosome. These factors, when bound, may impart different stability to the particle either by allowing it to easily open or by keeping it in a "closed" state. This almost uninvestigated function may control the accessibility of factors to sites hidden in nucleosomal DNA, and thus numerous processes that require such access.
The possible participation of Nap1 in the processes of transcription initiation and elongation also requires further attention, especially as far as such a role in vivo is concerned. On the other hand, the properties that Nap1 possesses as a histone chaperone point to a possible involvement in other processes that occur in the context of chromatin: replication, recombination, and repair. None of these potential roles have been studied.
To make things even more complicated, the numerous protein partners recently identified clearly indicate that Nap1 plays a multitude of additional roles that are not at all related to its histone chaperone characteristics. Exactly how Nap1 performs these additional roles is far from clear. Nap1 may be viewed as a juggler protein capable of doing many things at a time (Fig. 8
). Understanding how these apparently diverse and unrelated functions are channeled, coordinated, and regulated is an enormous challenge that will require considerable and concerted experimental effort, and a tremendous amount of thought.
|
|
ACKNOWLEDGMENTS |
|---|
Note added in proof: Nap1 was recently identified as a binding partner of both heterochromatin protein 2 (HP2) and nucleosome remodeling factor (NURF), an ISWI-dependent chromatin remodeling complex, in coimmunoprecipitation experiments in Drosophila embryo nuclear extracts. NURF, HP2, and Nap1 also coeluted during chromatographic fractionations of the proteins from these extracts, as confirmed by mass spectrometry and Western blot analysis. These interactions may point to an involvement for Nap1, possibly in cooperation with HP2 and/or NURF, in heterochromatin-induced gene silencing.
From Stephens, G. E., Xiao, H., Lankenau, D.-H., Wu, C., and Elgin, S. C. R. (2006) Heterochromatin protein 2 interacts with Nap-1 and NURF: a link between heterochromatin-induced gene silencing and the chromatin remodeling machinery in Drosophila. Biochemistry, 45, 14990–14999.
Received for publication September 15, 2006. Accepted for publication December 25, 2006.
|
REFERENCES |
|---|
is required for RNA polymerase II transcription on chromatin templates. Nature 413,435-438[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  S. Wang and L. Frappier Nucleosome Assembly Proteins Bind to Epstein-Barr Virus Nuclear Antigen 1 and Affect Its Functions in DNA Replication and Transcriptional Activation J. Virol., November 15, 2009; 83(22): 11704 - 11714. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J. Andrews, G. Downing, K. Brown, Y.-J. Park, and K. Luger A Thermodynamic Model for Nap1-Histone Interactions J. Biol. Chem., November 21, 2008; 283(47): 32412 - 32418. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. C. Del Rosario and L. F. Pemberton Nap1 Links Transcription Elongation, Chromatin Assembly, and Messenger RNP Complex Biogenesis Mol. Cell. Biol., April 1, 2008; 28(7): 2113 - 2124. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Clark, D. F. Alvarez, M. Alexeyev, J. A. C. King, L. Huang, M. C. Yoder, and T. Stevens Regulatory role for nucleosome assembly protein-1 in the proliferative and vasculogenic phenotype of pulmonary endothelium Am J Physiol Lung Cell Mol Physiol, March 1, 2008; 294(3): L431 - L439. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
TOP
ABSTRACT
GENERAL INTRODUCTION TO NAP1
