Molecular models for post-meiotic male genome reprogramming

Abstract

The molecular basis of post-meiotic male genome reorganization and compaction constitutes one of the last black boxes in modern biology. Although the successive transitions in DNA packaging have been well described, the molecular factors driving these near genome-wide reorganizations remain obscure. We have used a combination of different approaches aiming at the discovery of critical factors capable of directing the post-meiotic male genome reprogramming, which is now shedding new light on the nature of the fundamental mechanisms controlling post-meiotic histone replacement and genome compaction. Here we present a summary of these findings. The identification of the first factor capable of reading a precise combination of histone acetylation marks, BRDT, allowed highlighting a critical role for the genome-wide histone hyperacetylation that occurs before generalized histone replacement. In this context, the recent identification of a group of new histone variants capable of forming novel DNA packaging structures on specific regions during late spermatogenesis, when hyperacetylated histones are massively replaced in spermatids, also revealed the occurrence of a post-meiotic region-specific genome reprogramming. Additionally, the functional characterization of other molecular actors and chaperones in action in post-meiotic cells now allows one to describe the first general traits of the mechanisms underlying the structural transitions taking place during the post-meiotic reorganization and epigenetic reprogramming of the male genome.

Keywords:

• acetylation
• chromatin
• epigenetics

Introduction

One of the unknown biological processes remains the molecular basis of post-meiotic haploid genome reprogramming [Rousseaux et al. 2008]. Our lack of knowledge of this phenomenon includes sporulation in lower eukaryotes and pollen formation in plants [Govin and Berger 2009], as well as mammalian spermatogenesis, all directing a spectacular genome compaction. Simple and fundamental questions regarding the molecular basis of genome compaction and reorganization, i.e., the fate of histones, are completely unanswered [Gaucher et al. 2010].

In recent years important efforts have been employed to identify and functionally characterize critical elements in action during post-meiotic phases of spermatogenesis, and to propose the first molecular models underlying haploid male genome reprogramming. Indeed, facing the unknown nature of the mechanisms controlling post-meiotic genome reprogramming, approaches based on factor identification and their functional studies appeared to be very promising. In our group these approaches included: i) a large-scale identification of factors through proteomic approaches using various elongating/condensing spermatid extracts, at the stages when the major genome reorganization occur and ii) comprehensive in silico analyses of testis-specific transcripts encoding chromatin-related proteins (harboring known chromatin related domains such as bromo, chromo, etc.) (for detailed information see [Rousseaux and Ferro 2009]). Based on these factor identification strategies and the subsequent functional analyses, the first molecular models for post-meiotic male genome reprogramming could be proposed.

Chromatin acetylation-dependent genome reprogramming

Employing the early detailed analysis of histone acetylation during mouse and human spermatogenesis [Faure et al. 2003; Hazzouri et al. 2000] a hierarchy could be established in the lists of factors obtained through proteomic and in silico approaches [Rousseaux and Ferro 2009] for subsequent functional studies. Taking into account the early investigations reporting a massive histone hyperacetylation in spermatogenic cells before histone replacement in different species [Govin et al. 2004], our investigations started by a detailed analysis of histone acetylation during mouse spermatogenesis, which had remained uncharacterised. This founding work allowed characterizing the wave of global hyperacetylation affecting histones just before their replacement by transition proteins (TPs). This work also showed that, after this genome-wide acetylation, the disappearance of acetylated histones is not homogeneous throughout the whole genome, but that specific acetylated genomic domains persist even when most of the histones are replaced by TPs [Govin et al. 2007; Hazzouri et al. 2000]. Figure 1 shows the occurrence of a massive histone hyperacetylation simultaneously with a dramatic decompaction of pericentric regions revealed by a major satellite (fluorescence in situ hybridization, FISH). Interestingly, and unexpectedly, these heterochromatin regions retain nucleosomes for a longer period compared to the other genomic region (Fig. 1).

Figure 1.  Post-meiotic histone hyperacetylation triggers the subsequent male genome reorganizations. A wave of genome-wide histone hyperacetylation occurs at the beginning of the elongation of spermatids just before histone replacement by transition proteins. BRDT was recently identified as the first factor capable of binding to hyperacetylated histone H4. Interestingly in the absence of functional double bromodomain testis-specific protein (BRDT), the spermatogenesis defects first appear before spermatid's condensation. The wave of histone hyperacetylation is also associated with specific reprogramming of pericentric regions (revealed by a fluorescence in situ hybridization (FISH) approach using the major and minor satellite probes, as indicated), which lose their heterochromatic nature and gain histone acetylation. Specific late expressing histone variants organize these regions inducing the formation of new DNA packaging structures. These histones remain associated with the male genome in mature spermatozoa and appear as good candidates for transmitting male-specific epigenetic information to the egg after fertilization.

All of the subsequent functional studies and the choice to focus on the most relevant factors were then based on these observations. For example, the study of bromodomain-containing proteins, potentially capable of interacting with acetylated chromatin, was privileged, and the nature of the heterogeneous genome organizations (nucleosomal versus non nucleosomal genome organization) that is established during the histone hyperacetylation phase in elongating spermatids, was unraveled [Govin et al. 2007]. These characteristics are of potentially high importance since they may have a decisive role in the establishment of the specific genomic organization recently described in mature spermatozoa [Hammoud et al. 2009; Arpanahi et al. 2009].

Four factors, including two containing bromodomains, potentially acting on acetylated chromatin or controlling histone hyperacetylation, were selected for functional studies in our laboratory. At the time of identification, they were all of unknown function [Caron et al. 2003; Pivot-Pajot et al. 2003; Caron et al. unpublished].

Here, we will briefly develop our functional data on one of these factors, BRDT, a double bromodomain testis-specific protein. Indeed, our investigations now indicate that BRDT plays a major role in controlling chromatin acetylation-dependent events in spermatids. Our early studies indicated that the protein has two functional bromodomains and possesses the extraordinary capability of specifically condensing hyperactylated chromatin [Pivot-Pajot et al. 2003]. A multidisciplinary approach was then undertaken to discern its function in spermatids. These studies identified BRDT as the first factor specific for combinatorial histone mark reading. Indeed, its first bromodomain (BD1) is the only known BD, which requires simultaneous modifications (acetylation) of both lysines 5 and 8 of H4 for high affinity binding, whereas its BD2 specifically binds to H3 K18ac [Moriniere et al. 2009]. Therefore, efficient binding of BRDT needs simultaneous acetylation of three lysines on two histone tails. Moreover, since the simultaneous acetylation of K5/K8 specifically occurs in hyperacetylated H4, the requirement of acetylation of both H4K5 and K8 for BD1 recognition also demonstrates that this protein is the first known factor specific for binding tetra-acetylated histone H4. BRDT is therefore the perfect candidate which could act on hyperacetylated chromatin and mediate the subsequent events in elongating spermatids. Accordingly, it has recently been published that mice expressing a mutated from of BRDT lacking its BD1 are sterile and show post-meiotic defects [Shang et al. 2007].

Based on the observation of genomic ‘islands’ containing surviving acetylated nucleosomes in elongating/condensing spermatids, we developed an approach to characterize the nature of the histones involved. This work led to the discovery of a group of yet unknown histone H2A and H2B variants, which we named H2AL1, H2AL2, H2AL3, and H2BL1 and H2BL2. Our detailed studies of H2AL1/L2 showed that they are specifically expressed in late developing post-meiotic male germ cells, in condensing spermatids, at the same time as TPs, and that they remain present later in mature spermatozoa. We also observed that, after the wave of histone hyperacetylation, pericentric regions undergo a very unique reprogramming process, which ends by the assembly of these histone variants in these regions [Govin et al. 2007]. Therefore these histone variants constitute important characterized elements of the male-specific epigenome that could be transmitted to the egg [Wu et al. 2008].

Along this line of investigation, we showed that during spermatogenesis, the paternally and maternally methylated imprinting control regions (ICRs) carry different histone modifications during the stages that precede the global histone-to-protamine exchange. These chromatin differences could influence the way ICRs are assembled into specific structures in late spermatogenesis, and may thus influence post-fertilization events [Delaval et al. 2007].

Finally, taking advantage of conceptual similarities between spermatogenesis and yeast sporulation, we have also developed a pioneering work to unravel the fundamental and conserved mechanisms underlying post-meiotic genome compaction. We have investigated the state of H4 serine 1 (S1) phosphorylation during mouse spermatogenesis and at the same time this modification was investigated during yeast sporulation [Krishnamoorthy et al. 2006]. These investigations suggest a critical role for H4S1 phosphorylation in genome compaction.

Other factors identified through the analysis of specific extracts' proteomes, allowed us to discern some of the unknowns of post-meiotic genome structural transitions. Several chaperones in action in condensing spermatids were identified and among them we showed that HSP70.2 (HSPA2), a testis-specific member of heat shock proteins, associates with TPs in post-meiotic cells and is probably involved in their assembly/removal [Govin et al. 2006]. Within other chaperons present in elongating spermatids we also identified NAP1L4 as the chaperon of H2AL1/L2 [Govin et al. unpublished].

The work described above has allowed researchers to functionally dissect one of the less explored phenomenon in modern biology, paving the way to develop an active translational research program. These efforts have led so far to the identification of a mutation in Aurora C as an important genetic component of male infertility [Dieterich et al. 2007], and resulted in the identification of an amino acid change in BRDT responsible for male infertility in European and North African populations [Rousseaux et al. unpublished].

Conclusions

In the past years our various approaches for the identification of the most relevant factors and their subsequent functional characterization allowed to establish a list of potentially critical factors involved in the reorganization and epigenetic reprogramming of the mouse male genome.

Despite an important body of information on the biochemical and structural properties of the identified factors, these studies are now reaching limitations since no convenient in vitro model for spermatogenesis is available. In order to study these factors in a real physiological setting, there is a need to develop specific mouse models. Using a knock-in approach, four embryonic stem (ES) cell lines were established, each expressing one of our tap-tagged factors. These include BRDT, a new bromodomain and AAA ATPase factor, the testis-specific H2B variant tH2B, and one of the testis-specific H2A variants discovered in our laboratory, H2AL2.

These models will help us to precisely determine the function of these critical factors in terms of interacting factors and the associated genomic regions. All of these efforts will hopefully lead to the establishment of the first molecular models for post-meiotic male genome reprogramming.

Declaration of Interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Abbreviations

BRDT:	=	a double bromodomain testis-specific protein
TPs:	=	transition proteins
FISH:	=	fluorescence in situ hybridization
BD1:	=	first bromodomain
ICRs:	=	imprinting control regions
S1:	=	serine 1.