Nejire/dCBP-mediated histone H3 acetylation during spermatogenesis is essential for male fertility in Drosophila melanogaster

Spermatogenesis in many species including Drosophila melanogaster is accompanied by major reorganisation of chromatin in post-meiotic stages, involving a nearly genome-wide displacement of histones by protamines, Mst77F and Protamine-like 99C. A proposed prerequisite for the histone-to-protamine transition is massive histone H4 hyper-acetylation prior to the switch. Here, we investigated the pattern of histone H3 lysine acetylation and general lysine crotonylation in D. melanogaster spermiogenesis to elucidate a possible role of these marks in chromatin reorganisation. Lysine crotonylation was strongest prior to remodelling and the deposition of this mark depended on the acetylation status of the spermatid chromatin. In contrast to H4 acetylation, individual H3 acetylation marks displayed surprisingly distinct patterns during the histone-to-protamine transition. We observed that Nejire, a histone acetyl transferase, is expressed during the time of histone-to-protamine transition. Nejire knock down led to strongly reduced fertility, which correlated with misshaped spermatid nuclei and a lack of mature sperm. protA and prtl99C transcript levels were reduced after knocking down Nejire. ProtB-eGFP, Mst77F-eGFP and Prtl99C-eGFP were synthesized at the late canoe stage, while histones were often not detectable. However, in some cysts histones persist in parallel to protamines. Therefore, we hypothesize that complete histone removal requires multiple histone modifications besides H3K18ac and H3K27ac. In summary, H3K18 and H3K27 acetylation during Drosophila spermatogenesis is dependent on Nejire or a yet uncharacterized acetyl transferase. We show that Nejire is required for male fertility since Nejire contributes to efficient transcription of protA and prtl99C, but not Mst77F, in spermatocytes, and to maturation of sperm.


Introduction
One of the most dramatic processes of chromatin remodelling occurs during the post-meiotic phase of spermatogenesis. The previously histone-based chromatin becomes a highly compact structure due to the nearly genome-wide replacement of histones, first by transition proteins, which are then replaced by small basic proteins called protamines. The general aspects and other characteristic features of this large-scale chromatin compaction are conserved from fly a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 modifying the ProtamineB-eGFP construct of [15]. For nejire functional studies the RNAi line v105115 (Vienna Drosophila Resource Center) was combined with ProtB-GFP, Mst77F-eGFP [15] and Prtl99C-eGFP [18] transgenic fly lines. For the nejire knock down, v105115 protB-GFP males were crossed against virgin females of bam-Gal4/bam-Gal4; Sp/CyO; bam-Gal4-VP16/ MKRS. Crossings for knock down experiments were maintained at 30˚C.

Culture of pupal testes and inhibitor treatment
Pupal testes cultures of transgenic flies carrying ProtB-mCherry were established as described in [23]. In short, pupal testes were dissected in Shields and Sang M3 insect culture medium (Sigma-Aldrich Cat#S8398) supplemented with 10% foetal bovine serum (heat inactivated, insect culture tested, Sigma-Aldrich Cat#F3018), 100 U/ml penicillin and 100 mg/ml streptomycin (Gibco-Invitrogen Cat No. 15140-148). The intact testes were transferred to a 25-well plate. For inhibitor treatment, generally, six pupal testes were used for each inhibitor and control per experiment. The experiments were repeated at least three times.
Testes were treated with anacardic acid (AA, Merck Biosciences; 28.69 mM in DMSO stock) and trichostatin A (Cell Signalling Tech.; 4 mM in ethanol stock) in appropriate dilution in culture medium. Control cultures with solvent alone were analysed in parallel. Cultures were incubated for 24 h at 25˚C.
Sterility tests. One adult male of each genotype was placed together with two wild-type virgin females in a vial for 3 days at 25˚C. After 3 days, the parental generation was removed from the vials. After 2 weeks, offspring in each vial were counted.

RNA isolation and qPCR
RNA from 300 Bam>>v105115 testes or v105115 undriven testes was extracted using TRIzol (Invitrogen). Extracted total RNA was treated with RNase-free Turbo DNase (Ambion) and purified using an RNeasy mini kit (Qiagen). 1 μg of DNAse treated and purified RNA was used for cDNA synthesis with the Transcriptor First Strand cDNA Synthesis kit (Roche) according to the manual in a 20 μl reaction. qPCR reactions were set up with 10 μl iTaq™ Universal SYBR1 Green Supermix (Bio-Rad), 4 μl ddH 2 O, 0.5 μl (10 μM) gene-specific primer A, 0.5 μl (10 μM) gene-specific primer B and 5 μl cDNA (1:25 dilution). qPCR was performed in three technical replicates on the Mx3000P qPCR. Ct values were normalized to the mRNA expression level of Rpl32. Mean value and SEM of three biological replicates was calculated using GraphPad PRISM version 5.03. For statistical analysis one-sample t-test with a hypothetical value of 1 was used.

Multiple histone H3 acetylations in early and late canoe stages of the histone-to-protamine transition phase of D. melanogaster spermatogenesis
In D. melanogaster, the histone-to-protamine transition takes place during the so-called canoe stage. Here, the early canoe stage is defined by the start of histone removal, and the late canoe stage by the start of protamine accumulation [7]. Hyper-acetylation of histone H4 is proposed to be a prerequisite for histone removal during spermiogenesis in flies and mammals. In addition, a simultaneous increase in histone H3 acetylation was shown in mammals [6]. However, little is known about a possible role of H3 acetylation in post-meiotic chromatin remodelling.
Histone H4 acetylation is continuously visible from the spermatocyte stage to the canoe stage, where H4 hyper-acetylation is evident. In contrast, the histone H3 acetylation displayed an unexpected specific pattern within the stages of spermiogenesis (Table 1). We aimed to identify the overall pattern of histone H3 acetylation in spermiogenesis of D. melanogaster using specific commercially available antibodies against acetylated lysine residues K4, K9, K14, K18, K23 and K27 in the N-terminal domain of histone H3 and K36, K56 and K64 in the globular core domain (summarized in Table 1). Of these antibodies anti-H3K4ac and anti-H3K64ac showed no specific signal in germ cells. H3K56 acetylation was only detectable in primary spermatocytes (data not shown). H3K27ac is present in spermatocytes and prominent later in early canoe stage nuclei (Table 1). In addition to acetylation in primary spermatocytes, the other antibodies displayed characteristic post-meiotic acetylation patterns that for the most part allowed us to group post-meiotic histone H3 lysine acetylation into three classes: (1) acetylation that occurs prior to the histone to protamine transition; (2) acetylation occurring during the time of the transition (late in the early canoe, and beginning of late canoe stage), and (3) acetylation that occurs when most of the chromatin is protamine based (summarized in Table 1; example shown in Fig 1). In the first group, histone H3K9, H3K18 and H3K27 acetylation ( Fig 1Aʹ) were detected prior to transition (Table 1, green). H3K9ac was strongest in young elongating and early canoe stage nuclei ( Table 1). The signal disappeared in conjunction with histone degradation. H3K18ac ( Fig 1Aʹ) and H3K27ac (Table 1, green) are remarkable because of their speckled patterns, which occurred only in early canoe stage nuclei.
The second group comprises H3K14ac ( Fig 1Bʹ) and H3K23ac (Table 1, blue); both showed a speckled pattern in early and late canoe stage nuclei. H3K36ac represents a third class, since it was exclusively found in late canoe stage nuclei, clearly visible as a speckled pattern (Fig 1Cʹ  and Table 1, orange).

The histone mark lysine crotonylation is detectable in spermatocytes and canoe stage nuclei
In mouse, a recently discovered new histone mark, lysine crotonylation (Kcr), was reported to show intense labelling during elongating spermatid steps 9-11 ([13] for review see [26]). The intense lysine crotonylation signal coincides with the genome-wide histone hyper-acetylation prior to the switch from histones to protamines shown by [6]. Since lysine crotonylation is an evolutionarily conserved histone mark in D. melanogaster and mouse, we investigated the distribution of this histone modification in Drosophila spermatogenesis by staining squashed preparations of adult testes using an anti-Kcr antibody. Kcr is detectable in primary spermatocyte nuclei over the chromosome regions (Fig 2, column 1, arrowheads). In post-meiotic stages, Kcr is restricted to canoe stage nuclei where the speckled signal is most prominent in early canoe stage nuclei (Fig 2, column 2; Table 1). Thus, we conclude that Kcr disappears in correlation with histone degradation.

Anacardic acid blocks H3 acetylation and lysine crotonylation whereas TSA induces premature H3 acetylation and lysine crotonylation at the round nuclei stage
As we have shown previously, histone acetylation is essential for the transition from histonebased chromatin to protamine-based chromatin in D. melanogaster spermiogenesis [23]. To gain further insights into the role of H3 acetylation and lysine crotonylation during the time of transition, we used our established culture system for pupal Drosophila testes monitored with RFP or GFP tagged protamines [27]. In these cultures, the histone-to-protamine switch takes place as within the intact fly. The time of transition is easily assessed since the first spermatid cysts switch to protamine-based chromatin between 24 and 48 hours after puparium formation (APF). These properties make cultured pupal testes a valuable test system to study the histone to protamine transition using inhibitor treatment.
To study the effect of inhibiting histone acetyl-transferase (HAT) and histone deacetylase (HDAC) on H3 lysine acetylation and lysine crotonylation specifically in the post-meiotic phase, we isolated pupal testes of ProtB-mCherry transgenic flies at around 24 h APF. To block histone acetylation, these testes were subjected in culture to 150 μM of the HAT inhibitor anacardic acid (AA) for about 24 hours. A known inhibitor of several HAT families [28]. AA neither induces apoptosis nor shows cytotoxic effects in cultured testes, but leads to a block in elongating spermatid differentiation [23]. To block HDAC activity we treated cultured testes for 24 h with 50 μM trichostatin A (TSA). TSA inhibits both class I and II HDACs [29,30] leading to premature H4 acetylation in Drosophila spermiogenesis, while showing no obvious effect on histone to protamine transition [23].
Since the behaviour of lysine crotonylation with respect to HATs and HDACs has only been described in cell culture experiments [13] we examined Kcr distribution after AA and TSA treatment in D. melanogaster spermatogenesis. In control testes, Kcr was visible at the early canoe stage up until the late canoe stage was reached (Fig 3Aʹ). At this stage protamines are also deposited in cultured testes, as monitored by ProtB-mCherry expressing post-transition spermatids, which were detectable after 24 h of incubation (data not shown).
In testes treated with AA no late canoe stage spermatids had developed (Fig 3B and 3Bʹ). As expected from our studies with H4 acetylation, both H3K14ac and H3K18ac are no longer detectable in AA treated testes (data not shown). For adding and removing Kcr on histones it has been suggested that the enzymes responsible are independent from those mediating lysine acetylation and deacetylation [13]. Thus, we expected no changes in Kcr distribution upon treatment with AA and TSA. Surprisingly, inhibition of HATs led to a complete loss of Kcr in cultured testes (Fig 3Bʹ) compared to control testes ( Fig 3Aʹ). TSA treatment increased acetylation at all stages where HDAC activity is required (not shown) and resulted in enhanced and premature Kcr addition in round and young elongating spermatids ( Fig 3Cʹ).
In summary, our data further strengthen the role of histone acetylation as being an essential feature in allowing D. melanogaster spermatids to develop further, including the transition from histone-to protamine-based chromatin. We identified Kcr as a new histone mark that depends on the acetylation status of the spermatid chromatin and probably plays a role during D. melanogaster spermiogenesis.

The histone acetylase Nejire/dCBP is detected in spermatocytes and in the late canoe stage
The extensive reorganisation of chromatin during spermiogenesis requires stage-specific patterns of histone lysine acetylation. Consequently, we searched for enzymes that set these marks and thereby play a putative role in chromatin remodelling and spermatogenesis. H3K18ac and H3K27ac are specifically detected in the early canoe stage (Table 1). These are marks set by the CREB-binding protein (CBP), a lysine acetyl transferase also known as Nejire/dCBP in Drosophila [31]. Therefore, we concentrated on analysing Nejire/dCBP. Squashed testes preparations stained with an antibody against Nejire/dCBP revealed expression of Nejire/dCBP in somatic cells of the testes (data not shown) and in the germ line ( Fig 4Aʹ). Nejire/dCBP was expressed in primary spermatocytes, then not detected in early post-meiotic stages, but reappeared in late canoe stage nuclei ( Fig 4Aʹ). Lysine acetylation mediated by Nejire/dCBP, i.e. H3K18ac and H3K27ac, was detected at the early canoe stage (Table 1). This seemed a contradiction to the visibility of Nejire/dCBP at the late canoe stage, were histones are not detectable anymore. We searched in FlyBase for other candidates with the ability to acetylate H3K18 and H3K27. The genome of D. melanogaster contains numerous genes with coding capacity for acetyl transferases. Corresponding transcripts were found in testes of adult males for many of them. For some of these acetyl transferases, the targeted histone or a defined lysine residue is known, but none of them was characterized as targeting H3K18 and H3K27 (S1 Table). Thus, we considered that Nejire/dCBP might be the responsible enzyme and only detectable in late canoe stage because Nejire/dCBP is present in chromatin remodelling complexes and not accessible to the antibody during the early canoe stage, while the N-terminal acetylation marks H3K18ac and H3K27ac are accessible to the antibody. At the late canoe stage, histones are not detectable anymore in the already highly condensed chromatin, thus at this time Nejire likely is in the nucleoplasm and therefore might accessible to the antibody at this time. Thus, we Inhibition of all acetylations by inhibitors led to the failure of protamination. Therefore, we asked whether knocking down Nejire/dCBP dependent H3 acetylation is sufficient to abolish protamination.
To obtain a functional insight into the in vivo role of Nejire/dCBP in protamination, we applied RNAi-mediated knock down specifically in male germ cells, since Nejire is also expressed in somatic cells of the testis. Knock down of Nejire/dCBP by the germ line limited driver line bamGal4 was efficient in spermatocytes and canoe stage spermatids (Fig 4Bʹ). In D. melanogaster, several sperm proteins act additively to compact the paternal genome [18]. Therefore, we analysed the expression of ProtA-eGFP, Mst77F-eGFP, Prtl99C-eGFP in the Nejire/dCBP knock down situation.
In an overview of whole mount testes, it was evident that eGFP-labelled sperm proteins are present in young males (0-6 h after hatching) and three day old males (72-78 h after hatching) (Fig 5). At both stages, the overall morphology of the testes and the presences of elongated flagella were similar in control and Nejire/dCBP knock down males. Spermatid nuclei expressed Mst77F-eGFP (Fig 5A-5D), ProtA-eGFP (Fig 5E-5H), ProtB-eGFP (Fig 5I-5L) and Prtl99C-eGFP (Fig 5M-5P). In the control (bam GAL4 driver line) eGFP-positive nuclei accumulate towards the seminal vesicles, and the synchronously developing spermatid nuclei of cysts were evident (arrow). However, in the Nejire/dCBP knock down situation additional elongated eGFP-positive nuclei (arrowhead) and small round or oval nuclei (double arrow) were scattered towards the hub region (asterisk). In the whole mounts this is particular visible in Fig 5J,5N and 5P. Therefore, we conclude that Nejire is essential for correct maturation of sperm nuclei.

Post-meiotic de novo H3K18 and H3K27 acetylation
We hypothesized that Nejire/dCBP acetylates the histone lysine residues shortly before the time of histone-to-protamine transition. Alternatively, acetylations might persist from the spermatocyte phase and become visible due to the enormous reduction in the nuclear volume [32]. To test this hypothesis of post-meiotic acetylation, we used our pupal testes culture system (as described above). The histone-to-protamine transition takes place between 50-60 h after meiotic divisions [23]. We treated pupal testes of ProtB-mCherry transgenic flies with the HAT inhibitor anacardic acid and the HDAC inhibitor trichostatin A. We chose an incubation time of 24 h to avoid effects of transcriptional regulation in the spermatocyte stage. Thus spermatids at the canoe stage, between 40 h and 60 h after meiosis, went through the spermatocyte stage before we took the testes into culture.
Immunofluorescence staining with anti-H3K18ac antibody revealed that in contrast to the untreated control testes (Fig 6Aʹ) H3K18ac was almost undetectable in post-meiotic stages in AA-treated testes (Fig 6Bʹ). Testes treated with TSA displayed an increase in the H3K18ac signal ( Fig 6Cʹ). Furthermore, H3K18ac was prematurely detected from young elongating spermatid stages onwards (Fig 6Cʹ, columns 3-5). This argues for the activity of a so far uncharacterized detected with an anti-H3K18ac antibody. Expression of the ProtB-mCherry fusion protein was used to distinguish early and late canoe stage spermatids (not shown). Treatment with AA leads to a complete absence of H3K18ac signal in post-meiotic spermatid stages (Bʹ, column 4). In comparison to untreated testes (Aʹ, columns 4 and 5) further spermatid differentiation is blocked. In primary spermatocytes H3K18ac was strongly reduced (Bʹ, column 1). Treatment with TSA induced premature H3K18ac appearance in young elongating spermatid nuclei (Cʹ, column 3). TSA also induced increased nuclear H3K18ac levels in primary spermatocytes (Cʹ, column 1). Scale bar: 10 μm.
https://doi.org/10.1371/journal.pone.0203622.g006 enzyme with the same target specificity as Nejire/dCBP. In this experiment differentiating spermatids arrested before or during the histone-to-protamine transition. This might be due to inhibition of more than H3 acetylations in this assay, for example H4 acetylation [23].
Based on these findings we propose that Nejire/dCBP-mediated H3K18 and H3K27 acetylation in spermatids is a specific post-meiotic activity.

Nejire/dCBP is essential for high levels of protB and prtl99C transcripts
When monitoring Nejire/dCBP (Fig 4Aʹ) and H3K18ac and H3K27ac (Table 1) during the spermatocyte highly active transcriptional phase, these signals were severely reduced after RNAimediated nejire knock down (Figs 4Bʹ and 7Bʹ; Panel Bʹ in S1 Fig). Mst77F, protA, protB and prtl99C genes are transcribed in spermatocytes and transcripts are under translational repression until the end of the canoe stage. The corresponding proteins accumulate during the late canoe stage in the spermatid nucleus [18,33]. It is likely that transcription of these genes depends on Nejire/dCBP. To test this, we isolated RNA from testes of the undriven RNAi line v105115 and bamGAL4 driven RNAi (v105115) to quantify nejire transcripts. We determined the relative transcript level of mst77F, protA, prtl-99C, nejire/dCBP and tbrd-1 as controls (Fig 8G).
In the knock down situation, nejire transcripts are reduced considerably. The residual nejire transcripts were likely due to Nejire expression in somatic cells of the testes, which are not affected by the bamGAL4 driver line. tBrd1 transcripts and proteins are limited to the spermatocyte stage [34]. We chose tbrd1 transcripts since Mst77F, prtl99C, protA and protB genes do not belong to tBRD1 target genes [35,36]. tbrd-1 transcript levels were not significantly changed upon nejire knock down. Mst77F transcript levels were not severely reduced in the knock down situation in contrast to protA and prtl99C transcripts, which were reduced to about 40% of the wild-type level (Fig 8G). We conclude that the transcript levels of protA and prtl99C depend directly or indirectly on Nejire/dCBP, and thus might arrest spermiogenesis due to insufficient synthesis of sperm chromatin components.

Nejire/dCBP is essential for efficient post-meiotic histone degradation and male fertility
We subsequently analysed squash preparations for ProtB-eGFP in the Nejire knock down situation with respect to Nejire/dCBP-dependent H3 acetylation, histone degradation and ProtB-eGFP deposition. Here, the characteristic H3K27 ( Fig 7Bʹ) and H3K18 (Panel Bʹ in S1 Fig) acetylation pattern was missing in spermatocytes and the early canoe stage. In spermatid bundles in cysts, histones were detectable at the early canoe stage (compare Fig 7Bʹʹ with Fig 7Aʹʹ) but undetectable at the late canoe stage, as in the wild-type situation. However, the spermatid nuclei at the early canoe stage seemed less compact (Fig 7B and 7Bʹʹ) than in wild-type (Fig 7A  and 7Aʹʹ).
The phenotypes presented in Fig 7 were most abundant; however, we also observed highly abnormally shaped spermatid nuclei (Fig 8B and 8Cʹ), in agreement with the aberrantly shaped nuclei in whole mount preparations (Fig 5). The shape of ProtB-eGFP positive nuclei was highly variable (Fig 8A-8Cʹ), although similar in the individual nuclei of one cyst. Some protamine positive and histone negative nuclei appeared in this assay (Fig 8Aʹ, arrow). Strikingly, in contrast to wild-type, several slim nuclei contained both histones and ProtB-eGFP ( Fig  8A-8Aʹ, arrowhead), a feature not observed in squash preparations of wild-type testes [23]. In particular, we found histones and ProtB-eGFP in early canoe stage-like nuclei (Fig 8B-8Bʹ) and in tid-shaped nuclei (Fig 8C-Cʹ).
Notably, we found no individualized sperm after Nejire/dCBP knock down in squash preparations (Fig 7). We analysed seminal vesicles of 3-day-old males in the Nejire knock down situation, and again, we could not detect residual sperm in their seminal vesicles (Fig 8E). In contrast, wild type seminal vesicles were full of protamine-eGFP positive sperm (Fig 8D). In agreement with these phenotypes, fertility tests demonstrated that knock down of Nejire led to male sterility (Fig 8F).
ProtB-GFP was used to visualize protamine expression (Aʹʹʹ, Bʹʹʹ). In the wild-type, H3K27ac characterizes the spermatocyte and early canoe stage (Aʹ), histones are detectable until the early canoe stage (Aʹʹ), ProtB-eGFP characterizes nuclei at the late canoe stage and in individualized sperm (Aʹʹʹ). (B) Knock down of Nejire led to a complete absence of individualized sperm and the H3K27 acetylation signal in both spermatocytes and post-meiotic spermatid stages (Bʹ). After nejire knock down histones persist until the early canoe stage and degrade as in wild-type (compare Bʹʹ to Aʹʹ). Residual ProtB-eGFP positive spermatid nuclei are observed at the late canoe stage (Bʹʹʹ). Scale bar: 5 µm.
https://doi.org/10.1371/journal.pone.0203622.g007 In summary, these data indicate that the efficiency of histone degradation and protamination is impaired by Nejire knock down and therefore depends directly or indirectly on H3 acetylation by Nejire/dCBP.

Discussion
Male germ cells in vertebrates and Drosophila undergo a striking compaction of their genome in the haploid phase leading to nearly complete replacement of the histone-based chromatin by protamines. For many years, scientists have been trying to unravel the molecular basis of post-meiotic genome reprogramming that ultimately leads to large-scale genome compaction by the deposition of highly basic DNA-interacting non-histone proteins such as transition proteins and protamines.

Histone acetylation is a prerequisite for histone crotonylation in early canoe stage spermatids
A characteristic feature of the post-meiotic chromatin reorganization phase is the deposition of stage-specific histone modifications in flies and mammals. Detailed descriptions of modifications such as acetylation, methylation, phosphorylation, ubiquitination and SUMOylation are available (reviewed in [3,37]. However, the most striking modification event linked to histone replacement in spermiogenesis is the replication and transcriptionindependent massive hyper-acetylation of the core histone H4 in early elongating spermatids (Fig 9). Since global histone hyper-acetylation by itself is not enough to mediate removal of histones we tried to identify further modifications playing a putative role in histone-to-protamine transition.
In addition to histone H4 hyper-acetylation and H3K79 methylation [11,38], we identified lysine crotonylation (Kcr) as another conserved histone modification that precedes removal of histones in the absence of global transcription in D. melanogaster spermiogenesis (Fig 9). Histone Kcr occurs in a hyper-crotonylation wave in elongating spermatids, coinciding with histone hyper-acetylation in mice [13]. Like histone acetylation and H3K79 methylation, Kcr is a mark of transcriptionally active chromatin in spermatocytes. Using the histone acetyl-transferase inhibitor anacardic acid in pharmacological assays with cultured testes revealed that lysine crotonylation is dependent on post-meiotic histone hyper-acetylation; thus, we assume that crotonylation marks were set de novo in spermatids. Since inhibition of acetylation also strongly reduced H3K79 methylation, these modifications likely depend on histone acetylation. These findings indicate that these histone modifications might act in concert to regulate chromatin remodelling during the histone-to-protamine switch.

A distinct pattern of H3 acetylation characterizes the histone-to-protamine transition phase
In addition, we analysed the distribution of histone H3 acetylation marks in Drosophila. In contrast to the conserved H4K5/8/12 acetylation, detectable from spermatocyte stages until histones are removed, H3 lysine acetylation was shown to display a versatile pattern in Drosophila spermiogenesis (Fig 9). H3K18, H3K27 and H3K9 acetylation are detectable in stages prior to chromatin remodelling. Here, the observed H3K18 and H3K9 pattern corresponds to the pattern of these acetylation marks in mammals [39]. In Drosophila H3K36 acetylation is detectable in late spermatids already loaded with protamines. The vast majority of histones is reported to be replaced by protamines at this stage, but using a specific antibody we showed that at least histone H3 is detectable for longer in canoe stage nuclei than the other core histones [7]. Of interest, a depletion of H3K9 acetylation in mice corresponded to a prolonged retention time of histone H3 on chromatin [12].
A fundamental question is the functional implications of the histone modifications that precede and accompany histone-to-protamine transition. Amalgamating all the stage-specific histone modifications and testis specific histone variants that have been reported, one might generate a testis-specific histone code directing chromatin compaction, histone removal and histone degradation by recruiting a specific machinery acting on modified histones. However, the nature of this histone code remains elusive. In mammals so far, histone hyper-acetylation Nejire/dCBP is essential for male fertility was hypothesized to lead to an open chromatin structure that facilitates and induces histone displacement.

Many histone modifications, HMG box proteins, chaperones, Nejire and ISWI characterize the switch from an open nucleosome-based arrangement to highly condensed sperm chromatin
In D. melanogaster, at least four proteins of sperm chromatin are known: the very similar ProtA and ProtB (also known as Mst35Ba/Bb) and Mst77F and Prtl99C (Fig 9). Biochemical evidence suggests that Mst77F multimerizes to compact chromatin [17]. Mst77F is deposited independently of protamines [40], while protamine deposition seems to depend on the presence of Mst77F [14]. We proposed that these proteins act additively to compact sperm chromatin [18].
Using a combination of in vitro cultures and inhibitor studies we show that crotonylation depends on previous acetylation, as already known for methylation of H3K79 by Grappa (Drosophila Dot1l) [11,38]. Strikingly, H3K79 methylation and the responsible Dot1-like methyl transferase Grappa characterize the early and late canoe stage [38]. We proposed that Grappa, in cooperation with Rtf1 and UbcD6, marks sites of DNA strand breaks, e.g. by H3K79 methylation, and recruits DNA repair proteins.
Since histone acetylation seemed crucial for progressing to histone-to protamine transition, we aimed to identify acetyl-transferases, in particular those expressed at the canoe stage. H3 acetylation appears later than H4 hyper-acetylation and very specific in the early canoe stage (H3K18ac and H3K27ac). H3K18ac and H3K27ac acetylation marks are set by the aceyl-tranferease Nejire/dCBP.
Here, we present evidence that H3K18 and H3K27 acetylation depends on Nejire. Nejire is essential for fertility and responsible for efficient protamine mRNA synthesis. Hyperacetylation of H3 at K18 and K27 during the histone-to-protamine transition likely is responsible for efficient histone degradation, since inhibition of histone acetyltransferases in the post-meiotic phase abolished H3K18 and H3K27 acetylation. These data are in agreement with de novo acetylation by Nejire/dCBP, if the detection in the early canoe stage indeed failed because of accessibility failure for the antibody. Alternatively, we propose that a so far unknown acetyl transferase with the same target specificity exists and that Nejire/ dCBP targets other nuclear proteins during the canoe stage. While in vitro cultures of testes allowed us to selectively address the consequence of inhibition on post-meiotic stages, RNAi knock down experiments are only possible in the highly transcriptionally active spermatocyte stage. Thus, the distortions seen after knock down of ISWI [14] and Nejire might reflect in part a consequence of their involvement in transcriptional regulation in the spermatocyte phase.
Expression of Nejire during the late canoe stage and the loss of H3K18 and H3K27 acetylation after stage-specific inhibitor application led us to hypothesise that Nejire has dual functions in transcription in spermatocytes and H3 acetylation shortly before and during the histone-to-protamine transition. Nejire/dCBP contains a bromodomain, a protein motif known to specifically bind acetyl-lysine residues (reviewed in [43]. We postulate that Nejire/ dCBP is recruited via its bromodomain to already existing histone acetylations and then acetylates H3K18 and H3K27.

Nejire regulates transcript levels of Prtl99C and protamines as well as efficient histone degradation
Nejire/dCBP is already expressed in spermatocytes, the major transcriptionally active phase ( [44] for review see [45]) and later in addition during the phase of protamine loading. This might argue for a dual role of Nejire during spermatogenesis. The knock down of Nejire/dCBP was performed with the male germ line driver line bamGAL4, which is active in spermatogonia and spermatocytes. Unfortunately, expression of GAL4 specifically after meiosis is not possible since there is no major transcriptional activity during spermiogenesis. Thus, we cannot interfere with Nejire translation specifically in spermatids. Although knock down of Nejire/dCBP in primary spermatocytes does not lead to obvious defects at this stage, reduced levels of protamine mRNA and of protamine-like 99C mRNA were observed. This might at least in part explain the observed sterility.
The time selective blocking of acetyl-transferases by anacardic acid after meiosis impedes H3K18 and H3K27 acetylation in late post-meiotic stages. Thus, we hypothesize that these acetylations at this time might contribute to efficient histone-to-protamine transition. Indeed, we observed co-occurrence of histones and ProtB in abnormally shaped spermatid nuclei suggesting inefficient histone degradation. Of interest, depletion of H3K9 acetylation in mice caused a prolonged retention time of histone H3 on chromatin [12].
De novo H3K18 and H3K27 acetylation takes place in the haploid phase, and the corresponding acetyl-transferase Nejire/dCBP is essential for male fertility Enhanced deposition of post-translational histone modifications in elongating spermatids is suggested to be a consequence of prior events, such as the down-regulation of enzymes involved in their removal from histones, rather than the presence of specific enzymes generating these modifications (reviewed in [5]). Here, we show that in particular H3K18 and H3K27 acetylations appear de novo during the preparation and implementation of the histone-to-protamine transition during the canoe stage. Strikingly, these modifications depend directly or indirectly on the histone acetyl-transferase Nejire/dCBP. In mice, a partially depleted version of CBP led to fertile male mice [46]. In contrast, the efficient reduction of Nejire/dCBP by RNAi in male germ cells of Drosophila led to male sterility, underscoring the importance of this acetyl-transferase for spermatogenesis.

Conclusions
In our view, Nejire likely regulates H3K18 and H3K27 acetylation during transcription in spermatocytes. H3K18 and H3K27 are prominent also during the histone-protamine transition. We hypothesise that male sterility is directly and indirectly dependent on Nejire and caused by failure H3K18 and H3K27 acetylation in both stages.
In conclusion, with respect to the histone-to-protamine transition we propose that Nejire regulates transcript levels for sperm relevant proteins in spermatocytes and in directly or indirectly acetylation of H3K18 and H3K27 for fully efficient histone degradation. We propose that the numerous different histone modifications are required for histone degradation.