Vincent W. Aoki, Douglas T. Carrell

During spermiogenesis, the protamine proteins play an integral role in spermatid chromatin compaction. Recent research has focused on many facets of protamine biology, including protamine gene and protein structure/function relationships, mechanisms of protamine expression regulation and involvement of the protamines in male fertility. In this paper, we review our current understanding of the structure and function of the protamine-1 (P1) and protamine-2 (P2) proteins and genes, the expression and regulation of these genes and the relationship between the protamines and male fertility. In addition, we offer a brief outlook on future investigation into protamine proteins.

During spermiogenesis, spermatid chromatin undergoes substantial compaction. Testis-specific nuclear proteins, the transition proteins and protamines, are responsible for this chromatin condensation [1-5]. The first step in this process occurs in haploid round spermatids and involves replacement of somatic histones with the transition proteins (TP1 and TP2). Subsequently, in elongating spermatids, the protamines (P1 and P2) replace TP1 and TP2. The resulting chromatin is highly condensed and transcriptionally silent.

In the two decades following the elucidation of the protamine protein sequences by McKay et al [6, 7] numerous studies have focused on these sperm nuclear proteins. The structure of the P1/P2 proteins and the genes encoding them have been well described. In addition, recent studies have uncovered the details of the unique expression of P1 and P2, including the storage of protamine transcripts until translation occurs during spermatid elongation [5]. Recent research has focused on the regulatory mechanisms controlling this temporal regulation of P1 and P2 expression during spermatogenesis [5]. Protamine research has also entered the clinic and we are beginning to understand the link between the protamines and male fertility. Meanwhile, there is evidence that TP1 plays a role in DNA repair processes and, together with TP2, may be required for complete processing of the P2 protein [8, 9]. In this paper, we review our current knowledge of the structure and function of the P1/P2 proteins and genes encoding them, the expression and regulation of the P1 and P2 genes, and the relationship between the protamines and male fertility. In addition, we offer a brief outlook on future investigation into protamine proteins.

The P1/P2 genes and proteins are highly conserved in the sperm of all mammalian species. The haploid genome encodes a single copy of the human P1 and P2 genes mapped to chromosome 16p13.3 [10]. In addition, TP2 is mapped to the same locus on chromosome 16p 13.3 [11]. The P1-P2-TP2 locus spans a 28.5-kb region and is organized in a linear array, a structural feature affording concurrent expression of the P1, P2 and TP2 genes [12]. This multigenic locus, therefore, represents a single coordinately expressed chromatin domain. Additionally, the mammalian P1 and P2 genes contain only one single intron [13, 14].

The structure of the protamine genes plays a major role in their transcriptional regulation. First, the P1 and P2 genes are located in a large methylated domain in round spermatids, which facilitates nuclear matrix attachment and potentiation of the P1-P2-TP2 gene locus [15]. In fact, this P1-P2-TP2 multigenic locus is flanked by matrix attachment regions (MAR) that contain repetitive alanine (Alu) elements, which serve as sites of methylation [16]. In general, methylation silences gene expression while hypomethylation serves to derepress these genes by allowing chromatin to bind to the nuclear matrix, thereby maintaining a potentiated state [16]. In the case of the P1-P2-TP2 locus, gene potentiation occurs after binding of the chromatin to the nuclear matrix, this association being mediated by the MARs [16]. Interestingly, Schmid et al recently reported that MAR binding in the P1-P2-TP2 locus was independent of Alu methylation status, which indicated that the MAR attachment site was important for transcription but was not regulated by the state of methylation [16]. Second, all protamine genes contain a TATA-box, which facilitates binding of transcription factors to their promoters, thereby playing a major role in transcription initiation. Third, a cAMP-response element (CRE) is present in all protamine genes and is highly conserved in sequence and location residing from position -57 to -48 [17]. The CRE regulates transcription by binding various CRE proteins to this regulatory region [18-21]. Fourth, upstream regulatory sequences in the individual P1 and P2 promoters bind other trans-acting proteins, thereby directing transcriptional activation or suppression [22-25].

The human P1 and P2 genes encode a 50 amino-acid protein and a final processed protein of 57 amino-acids, respectively (Figure 1). Overall, there is approximately 50 % identity between human P1/P2. The ratio of P1:P2 varies in each species and in fertile humans is approximately 0.9 to 1.0 [26-28]. Corzett et al recently reported the P2 content to vary from 0 % to 80 % in a selected group of mammalian species [29]. However, within a given genus the P1:P2 stoichiometry is tightly regulated [29]. Both the P1 and P2 proteins are highly basic and contain copious amounts of arginine (~50 % of the total amino acid compliment) and cysteine (~10 %). Although the arginine content of both proteins is similar, P2 is slightly more basic due to its higher histidine and lysine content. In light of this sequence information, the structure/function relationship is readily apparent. Indeed, the basic character of the protamines facilitates DNA binding due to the strong intermolecular attraction between the positively charged protamine and the negatively charged DNA backbone. Furthermore, the presence of cysteine residues allows the formation of disulfide bridges, affording stability in the protamine molecular structure.

Recent studies using electron microscopy and scanning probe microscopy have elucidated protamine mediation of DNA condensation. Binding of protamines to the DNA induces coiling of the DNA into a toroidal or doughnut-shaped structure. These loop domains are very compact, only half the size of somatic cell histone loops, accounting for the 6-fold increase in sperm chromatin compaction [35]. The mechanism by which the protamines mediate this interaction is currently under debate [36]. P1 and P2 may bind to the major groove of the DNA [30], attach to both the minor and major grooves [37] or electrostatically bind to the surface of the DNA by interacting with phosphate residues [38].

P1 is present in all mammalian spermatozoa while P2 has been detected in spermatozoa of mouse hamster, vole, rat, stallion and man. Although the P2 protein is not present in some other mammals, such as the bull and boar, the gene encoding P2 is present and transcribed in these species. The absence of the protein is explained by translational regulatory mechanisms [39]. From an evolutionary standpoint, it is likely that mammals have inherited the P1 gene from a common ancestor since it is present in all species studied thus far. There are two possible explanations for the origin of the P2 gene. First, P2 may be derived from P1 followed by the divergent evolution of the two genes [40]. Second, it may be possible that P1 and P2 were inherited from a single common ancestor and that successive species have lost the ability to express P2 protein [13].

The protamines are expressed in the post-meiotic haploid spermatid. An interesting aspect of this expression is that transcription and translation of the proteins are temporarily uncoupled in the developing spermatid [41]. Because protamine-DNA binding results in chromatin condensation and consequent transcriptional cessation, the protein cannot be transcribed concomitantly at the time it is needed during spermiogenesis. Thus, all transcription must precede translation of the pro-tamines. Indeed, P1 and P2 mRNA have been identified in round spermatids while the proteins remain conspicuously absent at this stage [42, 43]. Translation of these mRNAs is delayed until the elongating stage of spermatid differentiation [44-46]. We should note that a small amount of mRNA is present in the mature human spermatozoa, which presumably represents remnant message that was never translated [47].

The temporal uncoupling of transcription and translation of the P1 and P2 genes is due to transcriptional and translational regulatory mechanisms [48, 49]. Transcriptional regulation occurs on three fronts: (1) DNA methylation, (2) binding of trans-acting factors to TATA box, CRE-box or other specific promoter DNA sequences of the protamine genes and (3) potentiation via chromatin association with the nuclear matrix (Table 1). The P1/P2 genes are associated with regions of increased methylation in round spermatids as compared with elongating spermatids, accounting for the increased transcriptional activity at that stage [43]. Transcriptional regulation via the TATA box may be mediated by a TATA-binding protein (TBP) recently identified in rodents [50]. TBP binding to the TATA box results in transcriptional activation of the P1/P2 genes [50]. CRE nuclear factors mediate transcriptional regulation via the adenylyl-cyclase pathway [48]. These nuclear factors include cAMP-response element binding protein (CREB) and cAMP-response element modulator protein (CREM) [48]. Following the typical cAMP-protein kinase A pathway, CREB and CREM are phosphorylated allowing them to bind to the CRE. This binding, in turn, activates transcription of the P1/P2 genes.

Table 1. Testis-specific regulators of P1 and P2 transcription and translation. *(+) Activation; (-) Repression.

Other trans-acting factors include protamine activating factor (PAF-1) and a testis-specific Y-box protein (Y-box protein p48/p52), which activate haploid cell P2 transcription by binding to specific regulatory sites located at positions -64/-48 and -84/-72, respectively [51, 52]. This P2 transcriptional activation was exemplified by studies, which demonstrated a 5-fold increase in P2 transcription when PAF-1 and Y-box protein p48/p52 bound to these regulatory sites [51, 52]. Furthermore, alterations to the gene sequences of these binding sites resulted in a substantial reduction in P2 transcription [51, 52]. In addition, there are a number of ubiquitous and testis-specific proteins, such as Tet-1 and nuclear factor I, which bind with the P1 promoter to activate P1 transcription [53, 54]. Thus far, no regulatory proteins have been identified which specifically function to inhibit transcription of the protamine genes. Although there are general modifiers of gene activity such as the Y-box proteins in which interactions with other tissue specific regulatory proteins determine their role in gene regulation [55].

Lastly, the potentiation of the P1-P2-TP2 locus may be mediated through association with the nuclear matrix [56]. Lee et al demonstrated aberrant premature expression of the P1-P2-TP2 locus causes developmental arrest and cell death [57]. It makes sense, therefore, that in non-expressing cells the chromatin is maintained in a transcriptionally silent closed conformation. Then, during spermatogenesis, the chromatin is selectively opened in specific domains to express those genes needed during differentiation [56]. As mentioned above, in the P1-P2-TP2 locus, gene potentiation occurs after binding of the chromatin to the nuclear matrix, this association being mediated by the MARs [16]. Although these nuclear matrix attachment regions do not appear to be regulated by methylation status, their attachment to the nuclear matrix does afford potentiation of the P1-P2-TP2 locus [16].

Thus, protamine gene regulation can be integrated into a three-step process consistent with the recently emerging model for gene expression [56]. First, potentiation of the protamine gene cluster occurs either by hypomethylation of the region or MAR binding to the nuclear matrix thereby rendering the gene cluster in an open chromatin structure affording transcription factor binding. Second, initiation of transcription via binding of trans-acting factors to TATA box, CRE-box or other specific promoter DNA sequences of the protamine genes. Lastly, an elongation phase in which the P1, P2, and TP2 transcripts are produced paving the way for translation.

Indeed, translational regulation is one of the most important aspects of protamine biology. In fact, if protamine transcription and translation is allowed to occur concurrently, the DNA undergoes precocious compaction and sperm development is arrested [57]. It is the mechanism of translational regulation that is responsible for the delay in P1/P2 protein appearance until the elongation phase of spermiogenesis. One mechanism to accomplish this temporal regulation of mRNAs is to keep the transcripts physically separated from the cellular translation machinery. In fact, translationally repressed mRNAs are sequestered in messenger ribonucleoprotein (mRNP) particles in chromatoid bodies near the sperm nucleus [58, 59]. Translational repression is evident during the round spermatid stage by the increased presence of these mRNPs [60].

Before being sequestered, however, the mRNAs undergo processing in the nucleus, which contributes to translation suppression. This processing includes a cleavage of the hnRNA primary transcript followed by the addition of a long polyadenylated (poly-A) tail, which serves to repress the mRNA from translation [5]. The poly-A mRNA is then stored in the mRNPs until deadenylation and shortening of the poly-A tail, at which time it is released from the mRNP and subsequently, undergoes translation [5]. Steger has proposed the regulation of adenylation may be under the control of cellular signaling cascades related to pituitary hormone control [5]. In addition, a poly-A binding protein (PABP) may have a dual role in the adenylation and consequent translation regulation of the protamines. First, by binding to the poly-A tail, PABP initially protects the mRNA from degradation thereby preserving the transcript until translation repression is removed [61]. Second, PABP may act as a repressor protein when it is bound to the end of the poly-A tail [61]. This repressor action is an indirect consequence to its protecting and maintaining the long poly-A tail. However, PABP then migrates upstream to the 3?UTR segment leaving the poly-A tail subject to degradation and subsequent translational derepression [61].

Sequence-specific RNA binding proteins also target the 3'UTR sequence of P1 and P2 mRNAs to mediate their translational regulation [62]. In general, it is thought that binding of these regulatory proteins to the protamine mRNA induces translational repression while modification of the mRNP complex results in release of translatable mRNA and subsequent translation derepression. Translational repression involving the 3'UTR sequences is accomplished mainly by the Y-box protein family [63]. These Y-box proteins contain a central cold shock domain that facilitates binding with nucleic acids. In general, phosphorylation of the Y-box protein is required for this binding to occur. Identified Y-box proteins which repress translation of the P1/P2 transcripts include the 48/52 kDa Y-box proteins (murine-MSY2, human-contrin and rat-YB2/RYB), the 18 kDa RNA-binding protein, the 26 kDa testis/brain RNA-binding protein (murine-TB-RBP, human-translin) and the 48/50 kDa RNA binding proteins [64-73].

A related protein in the rat, Prbp, has been shown to be required for translational activation of the P1 mRNA [74]. Prbp contains two copies of a double-stranded-RNA-binding domain and, in vitro, binds to the 3?UTR of P1 mRNA. The function of the protein in translation regulation was recently elucidated by Zhong et al [74]. This investigation showed that disruption of the gene encoding Prbp results in retention of translational repression of P1 mRNA. Consequently, the delayed replacement of the transition proteins by the protamines resulted in sterility and severely oligospermic mice [74]. Thus, the authors concluded that Prbp is required for the proper activation of repressed P1 mRNA. Curiously, this protein has been shown to bind to the 3?UTR of P1 mRNA, a well-defined characteristic of translational repressors noted above. In fact, a recombinant form of Prbp has been shown to repress translation of multiple mRNAs in wheat germ lysate [57]. Given this apparent discrepancy, a well-defined mechanism of Prbp action is very much needed but remains elusive. Indeed, such a study would be of great value in the light of a recently identified protein, TAR RNA binding protein (TRBP), which is the human homologue to Prbp and is expressed in spermatids at steps 3-4 of spermiogenesis, suggesting a potential role in protamine translation regulation [75].

The temporal expression of the protamine genes is indeed unique and extremely complex. Transcription occurs in round spermatids and involves potentiation of the P1-P2-TP2 chromatin domain, initiation via binding of trans-acting factors and elongation of the primary transcript. Subsequently, translational regulation accounts for the delay in protamine protein production and begins immediately during RNA processing with cleavage and polyadenylation of the mRNA. As the protamine mRNA enters the cytoplasm, it is translationally repressed via storage in mRNPs, binding by PABP and other RNA binding proteins, which target the 3?UTR sequence. Translation is then derepressed by covalent modification of the mRNPs, release of translatable mRNA and migration of the PABP to the 3?UTR, leaving the poly-A tail susceptible to degradation (Figure 2).

Because of the critical role normal protamine replacement plays in spermatid differentiation, one might expect aberrations in protamine expression or structure to lead to male infertility. Indeed, numerous reports have emerged in the last decade establishing a relationship between abnormal protamine expression and infertility. Premature translation of P1 has been reported to cause precocious nuclear condensation, which resulted in the arrest of spermatid differentiation in mice [57]. Abnormal P1:P2 ratios have been reported in the sperm of infertile human males, which indicate the relative amounts of each protamine is important for proper spermatid differentiation [27, 76-77]. In fact, there is evidence that the P1:P2 ratio is more important for male fertility than the absolute amount of protamines [58, 76, 78]. A recent report found P2 precursors to be present in the sperm of infertile males who had reduction in P2 levels [79]. Presumably, this indicates incomplete processing of the P2 protein. The importance of the P1/P2 ratio in spermatogenesis has recently been emphasized by a study, which showed that haploinsufficiency of P1/P2 causes severe infertility in mice [80]. In addition, the haploin-sufficiency of P2 has recently been reported to lead to sperm DNA damage and embryo death in mice [81]. Consistent with this data, Yebra et al reported the complete loss of P2 protein in a small proportion of infertile males [82]. Carrell and Liu later found that P2 was undetectable in 13 of 75 severely infertile patients analyzed prior to in vitro fertilization. Conversely, P2 was seen in all 50 donors of known fertility analyzed [27]. That study also showed that low P2 levels were generally associated with low sperm counts, motility and morphology. Additionally, Mengual et al recently reported marked increases in the P1/P2 ratios of oligozoospermic and asthenozoospermic patients as compared with fertile control patients [83]. Taken together these studies may indicate that the abnormal protamine content in sperm is a reflection of abnormal transcriptional or translational regulation of P1/P2 expression.

Functionally, it appears as if the protamines are required to impart zona pelucida binding and penetration abilities. This was exemplified by a study that showed destruction of the protamines inhibited sperm binding and penetration in the hamster egg penetration test [84]. Recent reports have shown that abnormal P2 is associated with diminished fertilization ability [85]. However, it does not appear as if normal protamine replacement is required for pronulcear formation, because ICSI with round spermatids has been shown to produce chromatin decondensation and pronucleus formation [84]. Additionally, patients without P2 have good success with ICSI during IVF cycles [85]. Furthermore, it has been demonstrated that destruction of the protamines actually increases sperm decondensation [84-85]. Taken together, these studies demonstrate that protamines are important components of spermatid differentiation and aberrations in the protamines are related to infertility and may reflect defects in spermiogenesis.

Thus far, the underlying causes of P1/P2 ratio deregulation in infertile males remains elusive. However, a number of attractive hypotheses may account for the induction of aberrant P1/P2 stoichiometry. First, mutations in either the P1 or P2 genes may play a role in P1/P2 deregulation. Although there have not been any substantial reports on mutations of the protamine genes in patients with specifically identified aberrations in the P1/P2 ratio, a recent study has identified single nucleotide polymorphisms in the P1 and P2 genes in a large group of both fertile and sterile male patients [86]. Second, gene mutations in any of the accessory proteins including TP1, TP2, Camk4 and Serine/arginine protein-specific kinase 1 may result in a breakdown in the normal histone-protamine replacement process and render an abnormal P1/P2 ratio. Third, in the absence of any gene mutations, an irregular P1/P2 ratio may reflect defects in translation regulation and, to a lesser extent transcription regulation through unfaithful function of accessory regulator proteins. Such defects could potentially disrupt the temporal expression of P1 and P2 and result in irregular P1/P2 ratios. Aberrant translation regulation is more likely to be involved in protamine expression regulation. Furthermore, the data showing disruption of regulatory proteins such as Prbp, which results in retention of translational repression and subsequent severe infertility, also support this hypothesis. Fourth, incomplete processing of the P2 protein could result in aberrant P1/P2 ratios. Indeed, there is evidence supporting this hypothesis, as mentioned above, a group of infertile males with reduced P2 levels also displayed an increase in P2 precursors [79]. Considering the importance of TP1 and TP2 in the processing of P2, this may be a reflection of TP deficiency or abnormal function. Finally, disruption of normal P1/P2 phosphorylation would result in incomplete protamine processing and unsuccessful binding to the DNA. Thus, improper signal transduction or malfunction in the appropriate kinases could potentially play a role in P1/P2 ratio deregulation.

As explained above, protamines are uniquely regulated during spermatogenesis. Proper transcription, translation and post-translational modifications are imperative to requisite function of chromatin compaction and transcriptional silencing during spermiogenesis.

Abnormal P2 expression has been shown in both mice and humans to lead to severe infertility. The identification of infertility patients with abnormal levels of P2 offers some interesting opportunities for future studies. First, it will be important to ascertain if an increase in the P1/P2 ratio of the proteins extracted from a pool of ejaculated spermatozoa reflects a general decrease in P2 in all sperm in the population or rather if P2 is diminished preferably in apoptotic or non-viable sperm. Second, the underlying causes of P2 deficiency can be evaluated, including the sequencing of relevant genes in affected patients. In addition to studying the P1, P2 and TP genes, it will be important to evaluate those genes that encode translation regulators, such as Prbp, which has been shown to lead to oligozoospermia when disrupted in mice. Third, temporal regulators may be responsible for abnormal P2 expression and should be further evaluated in patients with low P2 levels.

Finally, we suspect protamine production is a possible regulator or checkpoint of spermatogenesis. There are a number of studies supporting this hypothesis. First, haploinsufficiency of P1 and P2 leads to severely decreased spermatogenesis in mice [80]. Additionally, mice lacking P2 also show increased sperm DNA damage [81]. Interestingly, in the human, moderately diminished levels of P2 are associated with severely diminished sperm counts, motility and normal morphology [27]. Disruption of P1 and P2 expression, exemplified by a knockout of the translational regulator Prbp, also leads to severe infertility [74]. This novel hypothesis deserves further study, including the use of animal models to evaluate apoptosis during spermatogenesis in animals with targeted gene defects. Future studies will focus on that relationship, along with the potential causes of diminished protamine expression.