Histones poly ation

In order to determine the effect of core histone poly(ADP-ribosyl)ation on chromatin structure, we have incubated histone HI depleted chromatin with purified poly-(ADP-ribose) polymerase. This chromatin was then reconstituted with histone HI at a ratio of histone HI to DNA of 1. It was found that this chromatin did not recondense as the control chromatin did (A. Huletsky and G. de Murcia, unpublished observations). These results strongly suggest a close contact between histone H2B and histone HI and that modification of the N-terminal part of histone H2B, which was found to be poly(ADP-ribosyl)ated [32, 33], affects the integrity of the chromatin structure. In this process histone H2B would have an important role to play together with histone H1. 

Figure 5. Poly(ADP-ribosyl)ation-dependent histone shuttle on DNA. Step (1) shows a poly(ADP-ribose) polymerase molecule bound to chromatin. Step (2) auto(ADP-ri-bosyDation of the polymerase attracts histones from the DNA so they become noncovalently bound to the polymeric ADP-ribose chains attached to the polymerase. Step (3) indicates an extreme case where the local DNA has been completely denuded of histones. Step (4) upon degradation of the poly(ADP-ribose) by poly(ADP-ribose) glycohydrolase the histones reassociate with the DNA. (From Realini and Althaus, 1992).

Although a precise definition of the role of nuclear poly(ADP-ribosyl)ation is not available, the histone-shuttle mechanism proposed by Althaus and colleagues offers a possible unifying explanation of numerous experimental findings. While this model will come under further experimental scrutiny, the effects of ADP-ribo-sylating individual chromosomal proteins other than the polymerase itself (automodification) still needs to be elucidated. 

Realini C, Althaus FR. Histone shuttling by poly(ADP-ribosyl)ation. J Biol Chem 1992 267 18858-18865. 

As the in vivo introduction of new methyl groups on DNA involves the inhibition of poly(ADP-ribosyl)ation, which in turn can be correlated to a process of chromatin remodeling, parallel experiments were carried out, the only difference being that—for the in vitro reconsti-mtion of chromatin fibers—DNA sequences involved were either unmethylated or methylated with bacterial 5 I methyltransferase. The reconstitution of chromatin fibers was performed by adding either only the core histones or additionally H1 linker histone. 

Table 1 shows the requirements for the immobilized enzyme activity. DNA was almost absolutely required for the activity, as also for free (nonimmobilized) enzyme activities [2,3]. Histone HI, a potent activator of automodification of soluble enzyme [7], markedly inhibited the automodification of the immobilized enzyme in the presence of Mg", but slightly stimulated it in the absence of Mg". On the other hand, the histone, when included, served as a more efficient acceptor of poly(ADP-ribosyl)-ation than the gel-bound enzyme in the presence and absence of Mg. These results indicate that the histone served almost solely as an acceptor for poly(ADP-ribosyl)-ation by immobilized enzyme, which is in contrast to its actions as both an acceptor and an allosteric activator for free enzyme. 

Since the conditions described above, i.e., intermitotic arrest, proliferation, and DNA damage are associated with dramatic changes in nucleotide pools, we conducted a study to determine the effect of nucleotides and their components on protein acceptors for poly(ADP-ribosyl)ation [11, 12]. Figure 1 shows the effect of increasing concentrations of Ap4A on poly(ADP-ribosyl)ated proteins in human lymphocytes. In the absence of Ap4A, the autoradiograph shows that the most prominent ADP-ribosylated proteins are poly(ADP-ribose) polymerase at mol. wt. 116,000, histone H-1 at 32,000,... 

In order to elucidate the nature of the poly(ADP-ribose) polymerase interaction with the various components of chromatin, biochemical and morphological studies on various levels of organization of histone-DNA complexes were carried out. The poly-(ADP-ribosyl)ation of histone Hl-DNA complexes, core particles, chromatin depleted of histone HI, as well as native chromatin, were systematically studied using purified DNA-free calf thymus poly(ADP-ribose) polymerase in an in vitro reaction system. 

Biochemical and Morphological Study of the Poly( ADP-Ribosyl)ation of Native Chromatin, Histone HI Depleted Chromatin, and Core Particles... 

Furthermore, the acceptor proteins were analyzed by acid urea polyacrylamide gel electrophoresis (Fig. 2). Histones H2B, HI and protein A24 were found to be poly-(ADP-ribosyl)ated in native chromatin while in histone Hl-depleted chromatin and core particles it was found that histone H2B and protein A24 were poly(ADP-ribosyl)-ated. The presence of the hyper(ADP-ribosyl)ated forms of histone H2B on each of these various levels of chromatin structure have been confirmed by western blot analysis and by two-dimensional polyacrylamide gel electrophoresis (data not shown). The poly(ADP-ribosyl)ation of protein A24 have also been demonstrated by this last technique. 

Concomitantly, structures resulting from the poly(ADP-ribosyl)ation of native chromatin, chromatin-Hl and core particles were examined by electron microscopy. We found, as described earlier for pancreatic chromatin [6, 7], that calf thymus chromatin adopts a more relaxed conformation upon poly(ADP-ribosyl)ation by purified calf thymus poly(ADP-ribose) polymerase free of its DNA (Fig. 3a,d). It was also found that this chromatin exhibited a lower sedimentation velocity as compared to control chromatin [6]. And recently, it has been shown that DNA polymerase a activity is more than twofold higher in the presence of pancreatic polynucleosomes ADP-ribosylated as compared to control polynucleosomes [16]. In striking contrast, no ultrastructural effect was observed when chromatin depleted of histone HI was poly(ADP-ribosyl)ated (Fig. 3b, e). 

Interestingly poly(ADP-ribosyl)ation of chicken erythrocyte core particles resulted in the dissociation of nucleosomal DNA from the histone octamer (Fig. 3c,f). The dissociation coincided with the generation of the hyper(ADP-ribosyl)ated forms of histone H2B (Fig. 2). This result is noteworthy because on exposing hepatoma cells to DMS, histone H2B becomes hyper(ADP-ribosyl)ated [18]. The dissociation of nucleosomal DNA from the histone octamer after poly(ADP-ribosyl)ation explains why nucleosomal DNA becomes more accessible to micrococcal nuclease [19]. The increased accessibility of nucleosomal core DNA caused by poly(ADP-ribosyl)ation could explain in part the increased accessibility of DNA repair patches to micrococcal nuclease observed during the early phase of DNA repair [20]. 

Fig. 5. Electron microscopic visualization of poly(ADP-ribosyl)ated histone Hl-DNA complexes. Histone Hl-DNA complexes (H1/DNA 1/1) were formed and poly(ADP-ribosyl)ated as described in the legend of Fig. 4. Samples were then treated for electron microscopy as described in the legend of Fig. 3. a DNA, b histone Hl-DNA complexes (H1/DNA 1/1), c, d poly(ADP-ribosyl)ated histone Hl-DNA complexes. Notice the dissociation of histone Hl-DNA complexes after poly-(ADP-ribosyl)ation for 30 min at 200 ijM NAD. The bars indicate 1000 A...

This causes a relaxation of these structures (Fig. 5c,d) and this correlates well with the formation of the hyper(ADP-ribosyl)ated forms of histone HI (Fig. 4). It is tempting to suggest that the destabilization of these toroidal structures, resulting from the interaction between histone HI and DNA, is similar to what has been observed in the relaxation of the solenoid conformation by poly(ADP-ribose) polymerase [6, 7]. Our results strongly suggest that poly(ADP-ribosyl)ation of the chromatin and its subsequent structural relaxation are probably tightly coupled to the interaction between histone HI and DNA. 

This interaction between polymer and core particles and the fact that the structure of core particles can be dissociated by poly(ADP-ribosyl)ation leads us to suggest that during DNA replication poly(ADP-ribose) can serve as a deposition mechanism for histone octamer. 

Relationship Between Histone HI and Histone H2B Poly(ADP-Ribosyl)ation on the Chromatin Structure... 

Even though histone HI is mainly mono(ADP-ribosyl)ated during DNA repair [29], some poly(ADP-ribosyl)ation occurs as well. This pattern of modification suggests that... 

Indeed, we have found in chromatin depleted of histone HI [which has a structure similar to chromatin relaxed by poly(ADP-ribosyl)ation] that (1) histone H2B is easily poly(ADP-ribosyl)ated (Fig. 2) and (2) during in vitro polymer turnover on poly(ADP-ribosyl)ated nucleosomes, the polymer on histone H2B is much more resistant to glycohydrolase than the polymer present on hyper(ADP-ribosyl)ated histone HI [31]. 

Furthermore, some of our preliminary studies indicate that the turnover of poly-(ADP-ribosyl)ated proteins on polynucleosomes leads in time to an extensive modification of histone H2B. Thus, it is conceivable that during DNA repair in vivo, histone H2B modification results in destabilization of condensed chromatin and thereby permits increased accessibility to repair enzymes. Similarly, in active chromatin, where HI levels are greatly reduced, the damaged DNA might be more accessible to repair enzymes because of histone H2B hyper(ADP-ribosyl)ation. 

We have recently shown that, in vitro, poly(ADP-ribosyl)ation of nucleosomes either by purified poIy(ADP-ribose) polymerase [8] or by the endogenous chromatin bound enzyme [9] leads to relaxation of the chromatin superstructure throu histone HI modification. To test the reversibility of this phenomenon, we have used partially purified bull testis poly(ADP-ribose) glycohydrolase [10], which splits the ribosyl-ribose bond between two ADP-ribose units and produces ADP-ribose [11,12]. Changes of the nucleosome structure were examined by electron microscopy and by ultra-centrifugation, and the poly(ADP-ribosyl)ated histones were characterized. 

Thus, any model designed to explain the biological role of poly(ADPR) has to consider that both structural constituents like histone H2B as well as DNA-binding enzymes like topoisomerase I and poly(ADPR) synthase are modified by poly(ADPR). Assuming that poly(ADPR) synthase remains attached to the chromatin when activated by a DNA break, I favor the idea that poly(ADP-ribosyl)ation of the isolated acceptors is primarily a function of their accessibility to the synthase. The acceptor proteins may merely function as matrix to permit the accumulation of relatively large amounts of poly(ADPR) at distinct sites of the chromatin adjacent to the stimulating event. Thus, poly(ADPR), probably in the form of a three-dimensional network, may represent a specific tool to introduce changes into the chromatin structure. 

At present the events triggering and the mechanism controlling muscle differentiation are not well understood they may include changes in chromatin structure, gene amplification, and specific DNA and/or protein modification. One such modification, namely, poly(ADP-ribosyl)ation of nuclear proteins has recently attracted much attention and several studies have indicated that the enzyme poly(ADP-ribose) synthetase plays a role in cellular differentiation [6-9]. This enz)mie is a nuclear chromatin-associated protein which catalyzes covalent modification of both histone and nonhistone protein acceptors (for reviews see [10-13]). The synthetase is activated by DNA strand breaks and it has been suggested that DNA fragmentation and the consequent increase in poly(ADP-ribose) activity are obligatory events for chick muscle differentiation [6]. 

---