ANDROLOGY

ISSN: 2047-2919

REVIEW ARTICLE

Correspondence: Rafael Oliva, Human Genetics Research Group, IDIBAPS, Faculty of Medicine, University of Barcelona, Casanova 143, 08036 Barcelona, Spain. E-mail: [email protected]

Sperm nuclear proteome and its epigenetic potential 1,2

Keywords: chromatin, epigenetics, nuclear proteins, proteomics, spermatozoa

J. Castillo, 1,2,3A. Amaral and 1,2R. Oliva

1

Human Genetics Research Group, IDIBAPS, Faculty of Medicine, University of Barcelona, Biochemistry and Molecular Genetics Service, Hospital Clinic, Barcelona, Spain, and 3Biology of Reproduction and Stem Cell Group, Centre for Neuroscience and Cell Biology, University of Coimbra, Coimbra, Portugal

2

Received: 11-Oct-2013 Revised: 8-Nov-2013 Accepted: 11-Nov-2013 doi: 10.1111/j.2047-2927.2013.00170.x

SUMMARY The main function of the sperm cell is to transmit the paternal genetic message and epigenetic information to the embryo. Importantly, the majority of the genes in the sperm chromatin are highly condensed by protamines, whereas genes potentially needed in the initial stages of development are associated with histones, representing a form of epigenetic marking. However, so far little attention has been devoted to other sperm chromatin-associated proteins that, in addition to histones and protamines, may also have an epigenetic role. Therefore, with the goal of contributing to cover this subject we have compiled, reviewed and report a list of 581 chromatin or nuclear proteins described in the human sperm cell. Furthermore, we have analysed their Gene Ontology Biological Process enriched terms and have grouped them into different functional categories. Remarkably, we show that 56% of the sperm nuclear proteins have a potential epigenetic activity, being involved in at least one of the following functions: chromosome organization, chromatin organization, protein-DNA complex assembly, DNA packaging, gene expression, transcription, chromatin modification and histone modification. In addition, we have also included and compared the sperm cell proteomes of different model species, demonstrating the existence of common trends in the chromatin composition in the mammalian mature male gamete. Taken together, our analyses suggest that the mammalian sperm cell delivers to the offspring a rich combination of histone variants, transcription factors, chromatin-associated and chromatin-modifying proteins which have the potential to encode and transmit an extremely complex epigenetic information.

INTRODUCTION The mammalian spermatozoa is one of the most specialized cells, basically constituted by a markedly condensed nucleus, very little cytoplasm and an extremely long flagellum, with an overall hydrodynamic shape (Fig. 1A,B; Oliva & Dixon, 1991). The vast majority of the sperm chromatin is condensed by protamines in the form of highly compact toroidal structures, each one containing around 50 Kb of DNA (Fig. 1C,D; Balhorn, 2007; Oliva & Castillo, 2011a), although around 8% of the DNA remains organized by histones (Gatewood et al., 1987, 1990; Zalensky et al., 2002; Hammoud et al., 2009). The most classical and widespread idea about the function of the sperm cell is that its biological purpose is to transmit the paternal genetic message encoded in the DNA to the next generation. While the transmission of the paternal DNA sequence to the oocyte is beyond doubt, a very important sperm function, an increasing body of evidence is showing that the sperm cell also performs other © 2013 American Society of Andrology and European Academy of Andrology

essential functions. In addition to the inherited genetic information encoded in the DNA sequence, the epigenetic information is also crucial. The epigenetic information is constituted by DNA methylation, modifications of histones, presence of chromatin-associated proteins and RNAs, chromatin structure and chromosome territories in the nucleus (Mercer & Mattick, 2013; Rivera & Ren, 2013). It is now very well-known that proper DNA methylation imprints set during the male germline differentiation are essential for normal embryo development (McGrath & Solter, 1984; Trasler, 2009; Hammoud et al., 2010). In addition, the sperm cell delivers RNAs to the oocyte, some of which seem to be crucial for normal embryogenesis (Ostermeier et al., 2004; Liu et al., 2012; Jodar et al., 2013). Furthermore, some sperm proteins also play an essential role after fertilization. For example, in most mammals excluding mouse, the centrosome and thus the centrosomal proteins are paternally inherited (Chatzimeletiou et al., Andrology, 1–13

1

ANDROLOGY

J. Castillo, A. Amaral and R. Oliva

(A)

(B)

(C)

Histones and other chromatin associated proteins

Transcription factors, zinc fingers, chromatin modifying proteins, other proteins (D)

Nucleosomes Nuclear matrix (8% of chromatin) Attachment sites

Nucleosome (200 bp DNA)

Protamines

Figure 1 Different levels of structure of the sperm cell and its chromatin. (A) Frontal view of a phase-contrast photomicrograph of a human sperm cell. (B) Schematic drawing of the side view of a human sperm head, midpiece and part of the tail based on electron microscopy information. Note the extremely compact chromatin and the relative low proportion of cytoplasm. (C) Hypothetical model of the sperm chromatin composition and structure based on the known dimensions of the chromatin elements, on the nucleosomal distribution along sperm chromatin and on the detected presence of additional chromatin-associated proteins. On the right side an acidic polyacrylamide gel electrophoresis of human sperm proteins is shown demonstrating the presence of protamines and other less abundant proteins. (D) Dimensions of the nucleosomes and the protamine DNA toroids are shown together with the amount of DNA that they can pack (Arents et al., 1991; Balhorn, 2007).

Protamine-DNA toroid (50.000 bp DNA)

5.5 nm 25 nm 10 nm 50–70 nm

2008). Finally, it is not only the presence of specific DNA, RNA or protein components that have essential functions in the zygote, but also the structures determined by the combination of these different elements, providing additional layers of functionality and epigenetic information (Fig. 1C). Of relevance, genes important for the initial stages of development are organized into nucleosomes while the rest of the genes remain highly condensed by protamines, likely representing a layer of epigenetic information in the sperm chromatin (Fig. 1C; Arpanahi et al., 2009; Hammoud et al., 2009). For several years the knowledge of sperm nucleoproteins was mainly based on conventional methods which limited the identification to only the most abundant proteins in the sperm chromatin, that is protamines and histones (Oliva & Dixon, 1991; Bench et al., 1998; Oliva, 2006; Balhorn, 2007; Carrell et al., 2008). However, the largest contribution in the identification of the sperm nuclear landscape has occurred in recent years thanks to the application of advanced proteomic techniques based on mass spectrometry (MS) (de Mateo et al., 2013). Noteworthy, and as will be discussed in this review, besides histones and protamines, whole sperm proteomic analyses resulted in the identification of different transcription factors, zinc finger-containing proteins, chromatin-associated proteins and chromatin-modifying enzymes, which were postulated to have the potential to provide very important epigenetic information (Martinez-Heredia et al., 2006, 2008; Baker et al., 2007; de Mateo et al., 2007). The unexpected characterization of these proteins suggested the existence of other low abundant chromatin-associated proteins that would have escaped identification in whole cell analyses. 2

Andrology, 1–13

Therefore, subcellular proteomic approaches were successfully applied with the idea of enriching the proportion of low abundant proteins (de Mateo et al., 2011a; Amaral et al., 2013a, 2013b). Following this targeted approach, our group performed an isolation of nuclei from normozoospermic spermatozoa and identified 403 different proteins by MS, more than a half not previously described in the human sperm cell (de Mateo et al., 2011a). More recently, Baker and colleagues performed a proteomic study of isolated human sperm head fractions resulting in the identification of 703 proteins (Baker et al., 2013). On a technical note and as a result of the possibility of protein redistribution during nuclear purification, the proteins identified in isolated sperm nuclei are not all necessarily nuclear. On the other hand, whole sperm proteomic studies have detected nuclear proteins which have not been detected in isolated sperm nuclear proteome analyses. Therefore, a single catalogue containing all nuclear human sperm proteins was not yet available. Thus, in the present review we have compiled, analysed and reported a list of all chromatin-related proteins described so far in human spermatozoa (using either whole cell or head proteomic approaches) and discuss their potential implications. In addition, we have included and compared the nuclear proteomes of spermatozoa from different mammalian model species, demonstrating the existence of common trends in the chromatin composition in the mature sperm cell. The present review complements additional recent reviews covering different aspects of the sperm cell proteomics (Aitken and Baker, 2008; Oliva et al., 2008; Wu and Chu, 2008; Baker and Aitken, 2009; Oliva et al., 2009; Brewis and Gadella, 2010; Oliva and Castillo, 2011a; Baker © 2013 American Society of Andrology and European Academy of Andrology

ANDROLOGY

SPERM CHROMATIN PROTEINS

et al., 2012; Chocu et al., 2012; Dacheux et al., 2012; Dorus et al., 2012; Oliva, 2012; Oliva and Ballesca, 2012a; Rousseaux and Khochbin, 2012; Amaral et al., 2013b).

NUCLEAR PROTEINS IN HUMAN SPERMATOZOA To generate and analyse a catalogue of the human sperm nuclear proteins identified so far, we have used as a starting point the most complete list of proteins available for the human mature sperm cell (Amaral et al., 2013b). This list is constituted by 6198 different proteins and integrates all proteins reported in 30 different published proteomic studies performed using whole spermatozoa or subcellular proteomics, following a minimum of quality standards regarding the purity of the cells, the MS method and the protein identification criteria used (Table 1) (Amaral et al., 2013b). With the aim of identifying the subset of sperm proteins potentially involved in chromatin and nuclear function, we have selected proteins having at least one Gene Ontology (GO) cellular component term related to the nucleus using the web-based

bioinformatics tool DAVID v6.7 (Database for Annotation, Visualization and Integrated Discovery; http://david.abcc.ncifcrf. gov/) (Huang et al., 2009a, 2009b). To further identify additional nuclear proteins which could have escaped the GO search, we have completed the nucleoprotein list using a keyword based approach. Basically, we have manually selected those proteins in the combined human sperm proteome (Amaral et al., 2013b) that contained at least one of the following key words in their name: histone, bromodomain, transcription, DNA, nuclear/ nucleus/nucleo-, meiosis/meiotic, cell cycle/division, zinc finger, homeodomain, homeobox, chromodomain and plant homeodomain (PHD) Finger. Both lists of proteins (GO selected and manually selected) were then combined and the proteins were reviewed one at a time checking the functional descriptions available at the Uniprot database to exclude any proteins with no obvious nuclear function. At the end, we generated an accurate list of 581 human sperm nucleoproteins (Table S1). This list of proteins was then analysed using DAVID v6.7 to determine the nuclear biological processes enriched in the sperm cell.

Table 1 Summary of the sperm proteomic studies included in the analyses performed in the present review Reference

Human Amaral et al. (2013b)

Mouse Baker et al. (2008a)

Samples

Sample preparation/purification

Protein separation and digestion methods

Proteomic identification approach

Number of proteins reported

Compilation of 30 proteomic studies carried out using ejaculated spermatozoa. Main contributors: Baker et al. (2007, 2013), de Mateo et al. (2011a), Amaral et al. (2013a), Wang et al. (2013)

Whole and subcellular proteomics. Washing/Percoll Fractionation/Swim up/Puresperm

1D SDS-PAGE and in-gel digestion or in-solution digestion

LC-MS/MS

6198

Caudal epididymal spermatozoa

Maximum of 1 round cell for every 100 000 spermatozoa (microscopy analysis) Isolation of three distinct membrane raft subtypes. Washing by sequential differential centrifugation steps. Sperm purity confirmed by phase-contrast microscopy and by testing for the lack of an erythrocyte-specific membrane protein (TER 119) Sperm purity assessed using epifluorescence or differential interference contrast microscopy analysis 99% of sperm purity. Sonication

In-solution digestion and IPG strip peptide pre-fractionation

LC-MS/MS

858

1D SDS-PAGE, in-gel digestion

LC-MS/MS

190

In-solution digestion

HPLC-MS/MS

205

In-solution digestion, high-pH reversed-phase liquid chromatography fractionation 1D SDS-PAGE, in-gel digestion

LC-MS/MS

2850

LC-MS/MS

543

Asano et al. (2010)

Caudal epididymal spermatozoa

Dorus et al. (2010)

Caudal epididymal spermatozoa

Chauvin et al. (2012)

Caudal epididymal spermatozoa

Guyonnet et al. (2012)

Caput and caudal epididymal spermatozoa (only caudal proteins were included in the analysis)

Isolation of acrosomal matrix. Sperm filtering through a 10 lm diameter nylon mesh

Caudal epididymal spermatozoa Caudal epididymal Spermatozoa

At least 99.99% of sperm purity (microscopy analysis) Isolation of sperm heads by sonication. Washing with hypotonic buffer (0.45% NaCl)

In-solution digestion and IPG strip peptide pre-fractionation 1D SDS-PAGE, in-gel digestion

LC-MS/MS

829

LC-MS/MS

167

Caudal epididymal spermatozoa

Washing by repeated cycles of centrifugation with Ringer’s solution. Sperm purity assessed using DNAstaining and fluorescence microscopy

1D SDS-PAGE, in-gel digestion

LC-MS/MS

1247

Rat Baker et al. (2008b) Maselli et al. (2012)

Macaque Skerget et al. (2013)

1D SDS-PAGE, monodimensional sodium dodecyl sulphate polyacrylamide gel electrophoresis; IPG, immobilized pH gradient; LC-MS/MS, liquid chromatography tandem mass spectrometry.

© 2013 American Society of Andrology and European Academy of Andrology

Andrology, 1–13

3

ANDROLOGY

J. Castillo, A. Amaral and R. Oliva

DAVID v6.7 collects GO terms, among other data, and performs statistical analysis to identify overexpressed GO Biological Process terms in the list of gene products uploaded (Zhou et al., 2013). Remarkably, results from this analysis indicate that 56% of the sperm nuclear proteins have a potential epigenetic activity, being involved in at least one of the following GO terms: chromosome organization, chromatin organization, protein-DNA complex assembly, DNA packaging, gene expression, transcription, chromatin modification and histone modification. (p < 0.001; Fig. 2). Interestingly, the identified enriched biological processes included those involved in two main sperm epigenetic characteristics and past (or future) events: chromatin remodelling (i.e. nucleohistone to nucleoprotamine transition) and histone modifications (Miller et al., 2010; Carrell, 2012; Schagdarsurengin et al., 2012). Male germ cells experience massive chromatin remodelling events (Oliva & Castillo, 2011a, 2011b; Oliva & de Mateo, 2011). During spermiogenesis, a multistep replacement of histones by protamines takes place, resulting in a unique sperm chromatin

Figure 2 Human sperm nucleoprotein categories with a potential epigenetic function. (A) Abundance of the most representative enriched biological processes related to chromatin, based in Gene Ontology (GO) terms. (B) Number of proteins and corrected p-values (Benjamini) for each of the GO terms related to Biological Process. Related GO terms were joined. The gene-enrichment analysis significance was calculated by the web-based bioinformatic tool DAVID v6.7, using a modified Fisher Exact p-value. Benjamini correction was applied to express the maximum probability that the association is explained by chance alone. p < 0.001 was considered significant.

(A) Chromosome organization Chromatin organization DNA packaging Protein-DNA complex assembly Gene expression Transcription Chromatin modification Histone modification 0

(B)

4

50 100 150 200 250 Number of proteins (Count)

GO Biological ProcessTerm Chromosome organization Chromatin organization Chromatin assembly or disassembly Nucleosome organization Nucleosome assembly Protein-DNA complex assembly DNA packaging Gene expression Regulation of gene expression Transcription Regulation of transcription Transcription initiation from RNA polymerase II promoter RNA elongation Chromatin modification Chromatin remodeling Histone modification Histone H2A acetylation Histone H4 acetylation Histone acetylation

Andrology, 1–13

Count Benjamini 132 9.60E-78 101 4.60E-57 59 1.50E-47 52 4.90E-47 49 1.60E-45 51 3.00E-46 57 2.20E-47 212 4.40E-27 199 2.90E-22 159 3.10E-20 180 3.20E-19 12

2.80E-04

10 47 13 23 8 8 12

9.60E-04 3.00E-17 9.30E-06 7.80E-09 1.10E-06 6.00E-05 1.30E-05

structure with the major part of the DNA tightly packaged with protamines (Oliva & Dixon, 1991; Balhorn, 2007; Roque et al., € rndahl & Kvist, 2014; Gannon et al., 2014; Jodar and 2011; Bjo Oliva, 2014) (Fig. 1). A correct chromatin remodelling is required as protamination is important to condense and protect the DNA (Oliva, 2006; Torregrosa et al., 2006; Carrell et al., 2008; de Mateo et al., 2011b). In fact, alterations in the sperm protamine content seem to be related with DNA damage and male infertility (Oliva, 1995; Lewis & Aitken, 2005; de Mateo et al., 2009; Castillo et al., 2011; Simon et al., 2011; Kumar et al., 2012). Furthermore, protamine packaging is involved in gene silencing precluding transcriptional activity. Likewise, subsequent to fertilization, the highly compacted sperm chromatin is again remodelled: protamines are removed and replaced by maternally derived histones. Therefore, it seems coherent that in order to guarantee a correct chromatin structure, the mature human sperm cell is composed by proteins involved in chromatin organization, protein-DNA complex assembly, DNA packaging and chromatin modification (Fig. 2). Gene Ontology terms analysis of the nuclear-located sperm proteins also showed a significant enrichment of different biological processes related to histone acetylation (Fig. 2). Histone hyperacetylation is an essential step during the post-meiotic stage of male germ cell differentiation for the correct development and production of fertile spermatozoa (Oliva & Mezquita, 1982, 1986; Hazzouri et al., 2000; Faure et al., 2003; Oliva & Ballesca, 2012b). In addition, such histone modifications also have a role in gene regulation, as they modulate the binding of proteins involved in transcriptional activation (Rice & Allis, 2001). In fact, mature spermatozoa have a peculiar combination of histone modifications which constitute epigenetic marks that may be transmitted to the zygote after fertilization (Hammoud et al., 2011; de Mateo et al., 2011c). Finally, and although mature spermatozoa is considered transcriptionally and translationally inert, DAVID pointed to an enrichment in proteins involved in gene expression and transcription, including several transcription factors and protein domains such as zinc fingers or bromodomains (Fig. 2). The role of these events in spermatozoa is controversial and this field and the proteins involved will be extensively discussed in the next sections. To classify the 581 nuclear proteins, we divided them into the following families: (i) histones and histone modifiers, (ii) transcription factors, (iii) zinc fingers/bromodomains, (iv) DNArelated proteins, (v) ribonucleoproteins, (vi) meiotic/cell-cycle related, (vii) nuclear pore proteins and (viii) other nucleoproteins (Table 2; Table S1). The first four of these categories will be commented in the following sections. Histones and histone modifiers As previously mentioned, although most of the histones are replaced by protamines during spermiogenesis, around 8% of the human sperm chromatin (1–2% in mouse) remains attached to histones. Therefore, the identification of different histones in the combined human sperm nuclear proteome was totally expected (Table 2, Table S1; Fig. 1C). What is more surprising is the very large variety of histone members detected (46 different histones and histone variants) (Table S1). This number is remarkable considering that each of these histones may be differentially modified, defining the so-called histone code and © 2013 American Society of Andrology and European Academy of Andrology

ANDROLOGY

SPERM CHROMATIN PROTEINS Table 2 Number of proteins identified for each family of nucleoproteins described in human, mouse, rat and macaque sperm proteomes Human Zinc finger- and bromodomaincontaining proteins Histones and histone modifiers Transcription factors DNA-related proteins Ribonucleoproteins Proteins implicated in meiosis/ cell cycle Nuclear pore proteins Other nucleoproteins Total

Mouse

Rat

Macaque

89

14

2

17

67 78 51 42 34

31 18 33 35 15

5 1 5 1 1

9 3 18 3 5

28 192 581

16 37 199

4 3 22

3 7 65

creating an even more complex scenario (Jenuwein & Allis, 2001). As for other sperm proteins, sperm histones can be considered at two levels which are not necessarily mutually exclusive: histones with a role during spermatogenesis and histones putatively participating in early embryo development. Understanding sperm histone dynamics is particularly relevant as a growing body of evidence suggests that sperm nucleohistones transfer paternal epigenetic information to the embryo. The role of histones in chromatin remodelling during mammalian spermiogenesis is well documented (Oliva & Mezquita, 1982; De Vries et al., 2012). In fact, the displacement of histones by protamines is facilitated by the introduction of less-stable histone variants that may induce destabilization of the nucleosomes, as well as by a massive wave of histone acetylation (Oliva & Mezquita, 1982; Oliva et al., 1987, 1990; Govin et al., 2004; Gaucher et al., 2010). Interestingly, acetylation may be a mark for testicular histone degradation, as recent data suggest that testicular acetylated histones are degraded in proteasomes by means of a polyubiquitin-independent (but acetylation-dependent) mechanism, similar to the one operating in somatic cell DNA repair (Qian et al., 2013). Apart from facilitating histone replacement, histone variants (and their modifications) might also contribute to gene expression regulation during spermatogenesis. One of the detected and most well-characterized histone variants is the testis-specific histone H2B type 1-A (TH2B in the mouse) (Table S1). MS analysis of TH2B resulted in the detection of dynamic, stage-specific, post-translational modifications throughout spermatogenesis, which likely affect gene activity in the male germ line (Lu et al., 2009). Then again, the incorporation of this variant, which has the ability to form dimers with other testis-specific histones, seems to control the nucleohistone to nucleoprotamine transition (Govin et al., 2007; Montellier et al., 2013). In addition, the testis-specific H1 histone, which is phosphorylated at the C-terminal in elongating spermatids, and the X-chromosome encoded histone variant H2A.Bbd, are both present in the sperm proteome (Table S1), and may also participate in this transition (Rose et al., 2008; Ishibashi et al., 2010). On the other hand, the inactivation of the sex chromosomes seems to be mediated by the histone H2AX (Table S1) and its phosphorylated form (Fernandez-Capetillo et al., 2003). The significance of the histone code during spermatogenesis was also revealed by the inactivation of different enzymes involved in controlling histone modifications in mice models. To this extent, the inhibition of either specific histone deacetylases (or their recruiters), demethylases © 2013 American Society of Andrology and European Academy of Andrology

or of an enzyme involved in histone ubiquitination, resulted in different degrees of spermatogenesis defects and ultimately in male infertility (Roest et al., 1996; Fenic et al., 2004; Liu et al., 2010; Wasbrough et al., 2010). Similar to the mechanism operating in spermiogenesis (histone–protamine transition), the paternal chromatin remodelling occurring after fertilization (protamine-histone transition) may be mediated by histone hyperacetylation (van der Heijden et al., 2006). Unlike protamines, at least some (but possibly all) spermatozoa-derived histones are maintained, thus contributing to zygotic chromatin (van der Heijden et al., 2008; de Mateo et al., 2011c). The replacement of spermatozoa-derived protamines by maternally derived histones is a time frame when the male pronucleus is more accessible for the interaction with transcription factors. Consistent with this, the male pronucleus has a higher volume, shows higher transcription activity and contains increased concentrations of transcription factors compared with the female counterpart (Worrad et al., 1994; Aoki et al., 1997; Liu et al., 2005). Although the transcription machinery might be de novo synthesized via maternal RNAs (Aoki et al., 2003), there is the possibility that some of the transcription factors found in the sperm proteome (see below) may also participate here. Whatever the case may be, increasing body of evidence is showing that the expression of specific paternal genes (precisely the ones corresponding to the sperm nucleohistone fraction) may play a role in early embryo development. Furthermore, histone methylation patterns in spermatozoa may constitute marks of past and future programmes, with genes related to spermatogenesis (and presumably previously activated) associated with H3 lysine 4 methylation (H3K4me3; activating modification), and developmental promoters marked with H3K4me3, H3K4me2 (activating modifications) and H3 lysine 27 methylation (H3K27me3; repressive modification). Interestingly, there seems to be an overlap between bivalent genes (marked with both H3K4me3 and H3K27me3) in spermatozoa and embryonic stem cells, suggesting a role for sperm epigenetic marks in embryonic totipotency establishment (Hammoud et al., 2009). That active and repressive histone methylation patterns mark different male germ line promoters, with active ones mainly associated with genes relevant to spermatogenesis and bivalent ones with developmental regulators, was further confirmed by independent groups (Brykczynska et al., 2010; Erkek et al., 2013; Lesch et al., 2013). In addition, spermatozoa from some infertile men were shown to have randomly distributed histone retention, as well as a reduction in the amount of H3K4me3 or H3K27me3 in certain development transcription factors and imprinted genes (Hammoud et al., 2011). Taken together, the outcomes obtained so far suggest that histones, histone variants and their modifications are one of the main mediators of male germline epigenetic inheritance. The presence of so many additional different histones, histone variants (n = 46) and histone modifiers (n = 20) in mature human spermatozoa, as revealed by the analysis of the sperm nuclear proteome, is consistent with this idea (Table 2; Table S1). The field is now open to determine the specific gene associations and function of these additional sperm chromatin elements. However, in addition, the sperm cell also delivers to the oocyte many other nuclear proteins, some of which also have the potential to participate in early embryogenesis, as will be discussed in the following sections. Andrology, 1–13

5

J. Castillo, A. Amaral and R. Oliva

Transcription factors Gene transcription occurs from the earliest stages of spermatogenesis until the formation of elongated spermatids, just before the chromatin compaction by toroidal structures takes place (Oliva & Castillo, 2011b). In fact, at this post-meiotic stage of spermatogenesis, the haploid cell experiences the highest wave of transcriptional activity, which is regulated by different testis-specific transcription factors (Sassone-Corsi, 2002). After the replacement of nucleohistones by nucleoprotamines, however, sperm chromatin becomes condensed in such a way that the majority of the DNA is not accessible to interact with transcription factors (Fig. 1C). The mature male gamete is thus considered transcriptionally silent (Kierszenbaum & Tres, 1978; Stern et al., 1983; Oliva et al., 1988, 2008). Taking this into account, the identification of so many transcription factors in the sperm proteome and the enrichment in proteins involved in gene expression and transcription processes is puzzling (Fig. 2; Table 2). Which could be the role (if any) of this group of proteins in a cell where no transcription takes place? The simplest explanation is that the sperm transcription factors are spermatogenesis leftovers. In this case, the analysis of their expression in the mature sperm cell could still be useful as it may provide information as a proxy about the differentiation processes taking place in the testis, and therefore have potential clinical applications. Of note, spermatozoa possess several proteins involved in different steps of the RNA polymerase II transcription process (Table S1). These include, for instance, the TATA-box binding protein (TBP), a component of the macromolecular complex transcription factor IID (TFIID) that recognizes core promoter elements and is required for transcription initiation (Orphanides et al., 1996; Pittoggi et al., 2001). Overall, the amount of TBP transcripts in rodent adult testes, which seems to increase dramatically during late spermatogenesis, is more than 30 times higher than in any somatic tissue (Schmidt & Schibler, 1995). TBP is ubiquitously expressed in all stages of spermatogenesis, but in comparison with somatic cells, the overexpression of TBP proteins in the testis is lower than the observed for TBP mRNA. This, together with the fact that mature spermatozoa does not seem to contain sufficient proteomic machinery for transcription (Amaral et al., 2013b), suggests that the sperm TBP protein could be simply a remnant from previous spermatogenesis stages. The same may apply to all the other components of the RNA polymerase II transcription machinery identified in the mature sperm proteome. These include: members of the general transcription factors family, which constitute important components of the pre-initiation complex; subunits of mediators of RNA polymerase II transcription, which transmit positive and negative regulatory information, such as TFIID (Woychik & Hampsey, 2002); or elongated factors such as FACT (facilitates chromatin transcription) complex, constituted by SPT16 and SSRP1 subunits (Table S1). Alternatively, however, the sperm transcription factors could have a role in transcription regulation after fertilization, in early embryo development as it is commented below (Oliva et al., 2009; de Mateo et al., 2011a). It is also clear that the mature sperm cell contains a set of coding and non-coding RNAs (Miller, 2000; Ostermeier et al., 2005; Krawetz et al., 2011; Jodar et al., 2012, 2013). Although the functional significance of these RNAs is undetermined, roles in early embryonic development and epigenetic transgenerational 6

Andrology, 1–13

ANDROLOGY inherence have been suggested (Liu et al., 2012; Jodar et al., 2013). Indeed, the sperm cell delivers RNA molecules to the zygote (Ostermeier et al., 2004; Krawetz, 2005). Likewise, proteomic data have shown that the mature sperm cell contains many transcription factors known to be involved in differentiation and developmental processes (Table S1). Interestingly, proteins containing erythroblast transformation-specific domains, which are involved in developmental processes, have been identified in mature spermatozoa (Table S1) (Hsu & Schulz, 2000; Sharrocks, 2001). Also the T-box family (Tbx) members act as transcription activators recognizing tissue-specific gene promoters and have important roles in the regulation of transcription of genes involved in differentiation (Chapman et al., 1996; Papaioannou & Silver, 1998; Smith, 1999). Two members of Tbx family belong to the mature human sperm proteome, the TBX1 and the TBX5. The former although it is also called testis-specific component of the Tbx proteins, has been mainly studied in heart development context. Also noteworthy is the identification of transcription factors with one or more homeobox DNA-binding domains or homeodomains. The homeodomain is characterized by a 60 amino acids chain codified by a homeobox DNA conserved region (e.g. HOX cluster genes) which binds DNA or RNA. These proteins participate in transcriptional regulation in different developmental processes, such as cardiac development (Iroquois-class homeodomain protein IRX-4), retina formation (Homeobox protein SIX-5), neurogenesis (ventral anterior homeobox 1 and Homeobox protein cut-like 2) or normal growth and maturation of the lung (LIM-homeobox protein Lhx4), among others. One homeodomain-containing protein with highest levels of expression in mammalian testes is the putative homeodomain transcriptional factor 1 (PHTF1) (Manuel et al., 2000), which is one of the two members of the conserved Phtf gene family. PHTF1 was shown to be expressed in all spermatogenesis cell stages, with the exception of spermatozoa, and was suggested to be lost within the residual bodies (Oyhenart et al., 2003). It was thus striking to find PHTF1 in the mature sperm cell proteome and additional studies are needed to further clarify this issue. As a final note, and as previously mentioned, the sperm nucleohistone chromatin is enriched in developmental promoters and genes (e.g. Hox genes), as well as micro RNAs and imprinting genes as compared with the nucleoprotamine (Arpanahi et al., 2009; Hammoud et al., 2009). Thus, there is the possibility that the sperm transcription factors may also regulate the transcription of paternal nucleosome-associated genes soon after fertilization. As a first attempt to confirm this possibility, it would be of great interest to determine if certain transcription factors, as well as other proteins besides histones, preferentially occupy the sperm nucleohistone domains and specific genes critical for the development of the embryo, or, in contrast, have a random distribution throughout the chromatin. Zinc fingers- and bromodomain-containing proteins A high proportion of proteins identified in the human sperm nuclei contain zinc fingers, which are protein structural motifs involved in DNA/RNA and histone binding (Table 2). DNA-binding zinc finger domains confer roles in chromatin organization and/or gene expression regulation (Aasland et al., 1995; Ragvin et al., 2004). In fact, several sperm transcription factors contain this protein domain in their amino acid sequence. © 2013 American Society of Andrology and European Academy of Andrology

SPERM CHROMATIN PROTEINS

Importantly, various zinc fingers-containing proteins modify histones (i.e. histone acetyltransferases) or recognize histone covalent modifications, functioning as effectors or histone code ‘readers’ (Jenuwein & Allis, 2001). An interesting group of effector proteins found in the mature sperm cell proteome is that constituted by PHD Fingers. As previously explained, histone tail modifications originate epigenetic marks involved in gene expression activation or repression. PHD fingers are able to interact with and recognize several histone H3 modifications, that is H3K4me3, H3K9me3, H3K36me3, H3K9ac and H3K14ac, and are known to be essential for gene expression regulation, nucleosome remodelling and recombination (Musselman & Kutateladze, 2011). An example of a PHD finger protein found in the mature sperm cell is the nucleosome-remodelling factor subunit BPTF, which interacts with H3K4me3, a mark of transcription start sites, through its PHD finger domain. Interestingly, NURF301, the BPTF Drosophila orthologue, seems to be essential for maintaining HOX gene expression patterns during development (Li et al., 2006; Wysocka et al., 2006). This also depends on the chromodomain WDR5 (WD40-repeat protein), which is essential for HOX gene activation and vertebrate development (Wysocka et al., 2005). Chromodomains (chromatin organization modifier domains) are another remarkable group of effectors which can recognize methylated lysines (for reviews see Bannister et al., 2001; Ward, 2011). Bromodomains are also found in many chromatin- and transcription-related proteins acting as effector modules, but in this case by the selectively recognition of histone acetylations (Winston & Allis, 1999; Dyson et al., 2001). In fact, this protein motif is functionally linked to histone-acetyltransferase-associated transcriptional coactivators activity and is found in acetyllysine-binding domains (Dhalluin et al., 1999; Jenuwein & Allis, 2001). To this extent and because histone acetylation is indispensable in the chromatin remodelling process, bromodomains seem to be essential for the regulation of germ cell differentiation. The most studied bromodomain-containing protein in spermatozoa is bromodomain testis-specific protein (BRDT), a member of the subfamily of bromodomain-containing proteins BET. BRDT is required for the large scale acetylation-dependent chromatin reorganization through the binding to histone H4, essentially mediated by one of its two bromodomain motifs (Dhalluin et al., 1999; Pivot-Pajot et al., 2003; Govin et al., 2004; Moriniere et al., 2009; Dhar et al., 2012). In fact, genetic studies have demonstrated that the loss of this domain is enough to cause abnormal spermatids development and sterility in mice (Shang et al., 2007). Furthermore, the treatment of mice with JQ1, an inhibitor of the bromodomain and BET subfamily that blocks the acetylated H4 recognition through the occupancy of BRDT acetyl lysine binding cavity, leads to impaired spermatogenesis. Noteworthy and seeing that its action on male fertility seems to be reversible and its administration resulted in no secondary adverse effects in adult mice and their offspring, JQ1 was recently suggested as a novel male contraceptive molecule (Matzuk et al., 2012). On the other hand and as discussed for the other families of proteins, a putative role in the earliest stages of embryo development (especially during the paternal chromatin remodelling) can also be suggested for this and other sperm-bromodomaincontaining proteins. In fact, a potential role for BRDT during © 2013 American Society of Andrology and European Academy of Andrology

ANDROLOGY early embryo development has been proposed by others (Steilmann et al., 2010). Then again, and as indicated for transcription factors, it would be interesting to determine specifically which sperm DNA regions associate with these proteins and whether they are differentially distributed in the sperm chromatin domains. Other DNA-related proteins Mainly two subgroups of proteins were included in this group: enzymes responsible for DNA methylation regulation (DNA (cytosine-5) methyltransferases – DNMTs) and enzymes involved in DNA replication/repair and transcription (including DNA polymerases, helicases, replication licensing factors, mismatch repair proteins, topoisomerases and RNA polymerases) (Table S1). Reversible DNA methylation occurs on cytosine residues in cytosine-guanine dinucleotides (CpG), and is involved in gene expression regulation. Hypermethylation of promoter regions is related to gene transcription repression, whereas hypomethylation states are associated with transcriptional activation. Essentially, it is an epigenetic mark that promotes gene silencing and is critical to allele-specific expression of imprinted genes, X chromosome inactivation, cellular differentiation and development (for reviews see Hermann et al., 2004; Auclair & Weber, 2012). Accordingly, alterations in methylation patterns result in a spectrum of phenotypic defects. In mammals, de novo DNA methylation is catalysed by DNMT3A and DNMT3B and maintained mainly by DNMT1 (reviewed Bestor, 2000; Jurkowska et al., 2011). These enzymes might be crucial to male reproduction, as mice models with different members of the DNMT family disrupted in germ cells showed impaired spermatogenesis (Kaneda et al., 2004; Kato et al., 2007; La Salle et al., 2007; Takashima et al., 2009). Likewise, imbalances in DNA methylation during human spermatogenesis seem to alter gene expression and result in abnormal germ cell development and male infertility (Heyn et al., 2012). The testicular methylation pattern is distinct from that of somatic tissues, displaying more hypomethylated loci, but in non-CpG island sequences outside gene promoters (Oakes et al., 2007b). This unusual pattern begins to be established during foetal life, in embryonic germ cells, and carries on in the first stages of spermatogenesis, when de novo methylation and demethylation events occur (Oakes et al., 2007a). Actually, DNMTs transcripts and proteins were detected in all spermatogenic stages, with DNMT1 being the most abundant one (Marques et al., 2011). Thus, the presence of DNMT1 in the human sperm proteome was not unexpected, although its role in the male gamete (if any) remains unknown. The male gamete methylome is similar to that of embryonic stem cells and embryonic germ cells, showing a deep hypomethylation of developmental transcription gene promoters, which may contribute to the transcription of these genes in the early embryo (Weber et al., 2007; Arpanahi et al., 2009; Hammoud et al., 2009; Molaro et al., 2011). Noteworthy, sperm DNA methylation levels need to be tightly regulated, as alterations in the sperm methylome result in decreased pregnancy rate after in vitro fertilization (Benchaib et al., 2005). Epigenetic reprogramming in primordial germ cells is crucial for imprinting and the activation of parental-specific methylation patterns (Smallwood et al., 2011; Smallwood & Kelsey, 2012). Likewise, a second Andrology, 1–13

7

J. Castillo, A. Amaral and R. Oliva

genome-wide methylation reprogramming occurs in preimplantation embryos. After fertilization, and to remove gametespecific marks, there is a rapid and active genome-wide paternal DNA demethylation (as well as a passive maternal genome demethylation) of which only imprinted genes might escape (Reik et al., 2001; Abdalla et al., 2009; Smith et al., 2012). Of note, DNMT1 alone seems to be sufficient to maintain DNA methylation imprints during pre-implantation development (Hirasawa et al., 2008). The process of active demethylation of the paternal DNA is still poorly understood and the exact players in this process are undetermined. Interestingly, demethylation seems to be associated with DNA repair and thus the two processes may share their protein machinery (Razin et al., 1986; Weiss et al., 1996). Although the accepted paradigm is that all the proteins involved in paternal chromatin remodelling are maternally derived, there is the possibility that at least some of the sperm nuclear proteins may also play a role in this process. Although this idea will obviously need experimental confirmation, it is logical to think that, if not degraded once inside the oocyte, paternal-derived proteins could be preferentially used, as they might be physically closer (and thus more accessible) to interact with the paternal genome. On the other hand, enzymes involved in DNA replication and repair are relevant in the meiotic phase of spermatogenesis and may be present in mature spermatozoa as simple leftovers from past events. However, these enzymes could also participate in the reorganization of the paternal chromatin after gamete fusion. For instance, contrary to histone-bound DNA, protamine-bound DNA is not highly supercoiled, and topoisomerase II is believed to relieve supercoiling, thus facilitating histone displacement during spermiogenesis (Yap & Zhou, 2011). Indeed, it seems that histone hyperacetylation allows topoisomerase II to gain access to DNA and induce transient double-strand breaks in elongating spermatids (Laberge & Boissonneault, 2005). A reverse mechanism might operate during paternal chromatin decondensation, where the action of DNA topoisomerases may also be needed. Although DNA topoisomerases I and II isolated from nuclear extracts of human spermatozoa seem to have a lower enzymatic activity than their somatic counterparts (HarVardi et al., 2007), they might be transmitted to the oocyte and could thus play a role in paternal chromatin remodelling.

SPERM NUCLEAR PROTEOME IN MODEL SPECIES The sperm proteomic study of different model organisms is also providing catalogues of the corresponding sperm proteins resulting in very valuable information relevant to understand the different functions of the sperm cell and the conserved key elements. Mammalian sperm proteomic profiles are now available for mouse (Baker et al., 2008b; Nixon et al., 2009; Asano et al., 2010; Dorus et al., 2010; Chauvin et al., 2012; Guyonnet et al., 2012) (just to give some examples), rat (Baker et al., 2008a; Maselli et al., 2012), bull (Peddinti et al., 2008; Byrne et al., 2012; Soggiu et al., 2013), boar (Park et al., 2012) and macaque (Skerget et al., 2013). In addition, proteomic approaches have also been applied for the identification of sperm proteins in invertebrates and plants. For instance, protein catalogues of the fruit fly (Dorus et al., 2006; Yan et al., 2010), honey bee (Collins et al., 2006; Poland et al., 2011), cricket (Simmons et al., 2013), worm (Chu et al., 2006) and ascidian (Nakachi et al., 2011) 8

Andrology, 1–13

ANDROLOGY spermatozoa, as well as of rice and tomato pollens (Dai et al., 2006; Sheoran et al., 2007) have been described. With the aim of comparing the Homo sapiens sperm nuclear proteome with that of other mammalian species, we have used the published data to compile accurate lists of mature sperm nucleoproteins in other organisms. To avoid biased identifications and facilitate the comparative study with the compiled human protein list (Amaral et al., 2013b) (Table 1), mammalian proteomic catalogues available online were selected based on the following inclusion criteria: (i) spermatozoa was extracted from caudal epididymal regions (as our analysis is focused in the identification of nuclear proteins in mature sperm cells), (ii) a sample purification step was performed and/or sperm purity was checked and indicated, (iii) peptides were generated by either in-solution digestion or 1D SDS-PAGE in-gel digestion (i.e. 2D electrophoresis approaches were not included) and (iv) protein identification approaches relied on tandem MS (MS/MS) with at least two peptides per protein identified with a false discovery rate (FDR)