Both the sperm-associated and sperm-restricted GPx4 variants (mGPx4 and nGPx4) have been shown to be associated with intracellular proteins and to function as disulfide isomerases rather than as classical ROS-scavenging GPx. Concerning the sperm midpiece–located mGPx4, it has been estimated that this isoform constitutes up to 50% of the sperm midpiece protein content that embeds the helix of mitochondria (Ursini et al, 1999). For that reason it was proposed that mGPx4 is the selenoprotein of the sperm midpiece, a role given earlier to a protein called sperm mitochondria-associated cysteine-rich protein (Kleene, 1994). In the sperm midpiece, the mGPx4 protein is suggested to be more a structural protein than an active enzyme because it has been shown to have completely lost its solubility and its scavenging enzymatic properties (Ursini et al, 1999). It is, however, probable that the sperm midpiece–located mGPx4 is involved in local structural reorganization based on protein disulfide-bridging events. It has been shown that disulfide bonds in the late stages of spermatogenesis and during epididymal transit are important for several sperm structures (besides the nucleus), such as the plasma membrane, the midpiece, and the acrosome (Cummins et al, 1986; Mate et al, 1994; Francavilla et al, 1996; Mammoto et al, 1997; Sivashanmugam and Rajalakshmi, 1997). In the sperm midpiece, during spermiogenesis, it has been shown that mitochondria attach to outer dense fiber proteins of the axoneme and that disulfide bonds in several proteins are involved in this process. As a result, the spermatid cytoplasm is reduced and the sperm plasma membrane is connected to the sperm midpiece. In addition, it has been shown that the acrosome contains the greatest relative amount of disulfides, before the head and the tail in guinea pig spermatozoa (Huang et al, 1984), suggesting that there are regionalized disulfide-bridging events during sperm maturation. The group of M. Conrad (Schneider et al, 2009) generated a transgenic mouse model in which the mGPx4 was disrupted via the introduction of an in-frame translational stop into the mitochondrial leader sequence of mGPx4. The analysis of this mouse model reveals that mGPx4−/− mice are viable, contrary to the GPx4−/− mice (in which the somatic isoform[cGPx4] as well as the 2 sperm-specific variants [mGPx4 and snGPx4] are absent) that die during early embryogenesis (Imai et al, 2003; Yant et al, 2003). Interestingly, the mouse mGPx4−/− model showed male infertility associated with impaired sperm integrity. Essentially and quite logically, mGPx4−/− spermatozoa showed important structural abnormalities in the midpiece region, leading to an increase in bent flagella, sperm heads detached from the flagellum, abnormal distribution of mitochondria along the midpiece, and abnormal organization of the axoneme (Schneider et al, 2009). In addition, and confirming the disulfide-bridging function of the protein mGPx4, deficient spermatozoa exhibit a higher protein thiol content, and their phenotype resembles what occurs in severe selenodeficiency situations (Flohé, 2007; Shalini and Bansal, 2008). Also not surprisingly, sperm motility was significantly reduced in mGPx4−/− males. The authors showed that male infertility could be bypassed by ICSI, suggesting that the male gametes were unable to move properly as a consequence of sperm midpiece structural abnormalities and not because of their incapacity to initiate fertilization. Confirmation of these findings with regard to mGPx4 function was reported in Liang et al (2009) and Imai et al (2009). Using a different strategy, Liang et al (2009) generated transgenic mouse strains that carried mutations inhibiting the expression of either cytosolic or mitochondrial GPx4 and, consequently, overexpressing the other isoform. Their data confirmed that the mitochondrial GPx4 variant is testis- and male germ cell–specific. They also confirmed that when mGPx4 is not expressed, it leads to male infertility, essentially because of structural malformations of the sperm midpiece. The strategy used by Imai et al (2009) was to establish a spermatocyte-specific GPx4 knockout mouse via the Cre-loxP system. Again, this new transgenic mouse model showed oligoasthenozoospermia resulting in male infertility, confirming that a decrease in GPx4 activity in spermatozoa results in male infertility in mice.
The sperm nucleus–specific isoform of GPx4 (nGPx4) was shown to result from differential expression of its gene owing to the use of an alternative promoter located in the first intron (Moreno et al, 2003). This results in the expression of a GPx4 isoform having an N-terminus sequence rich in arginine residues, allowing its nuclear localization and binding to chromatin (Pfeifer et al, 2001). In sperm nuclei, the GPx4 variant has been proposed to act as a protamine thiol peroxidase responsible for stabilizing the condensed chromatin by cross-linking protamine disulfides (Pfeifer et al, 2001). Condensation of sperm chromatin is an essential process in sperm differentiation, which starts during postmeiotic spermatogenesis with the replacement of somatic histones by transition proteins and finally by protamines. It appears that the sperm DNA-packaging process is not totally completed when spermatozoa leave the testis and that it goes on in the early stages of epididymal maturation. During epididymal transit, oxidation of protamine thiols plays an important role in compacting sperm DNA further and also locking it in that highly condensed state. The cross-linking of protamine disulfides induced by ROS is comparable to GSH oxidation and peroxide reduction catalyzed by GPx. Therefore, it has been proposed that the spermnucleus GPx4 variant uses protamine cysteine residues as reducing partners and acts as a protamine thiol peroxidase (Pfeifer et al, 2001). For its activity, the nGPx4 isoform would not depend on GSH availability, which decreases significantly in late spermatogenesis and early maturation during epididymal transit. In agreement with that hypothesis is the observation that in selenium-deficient animals, in which the concentrations of selenium-dependent GPx such as nGPx4 are greatly reduced, nearly all sperm cells recovered from the vas deferens possess incompletely compacted nuclei. In addition, in vitro experiments have shown that dithiothreitol provokes rat sperm DNA decondensation, an effect that is restored by adding H2O2 (Pfeifer et al, 2001). Finally, it has been shown that the use of an nGPx4 inhibitor blocks the condensation of sperm DNA. Together, these data strongly support the idea that the sperm nucleus–located GPx4 variant is responsible for protamine disulfide bridging within the sperm nucleus. In 2005, Conrad et al generated a transgenic mouse model in which they specifically abolished the expression of the sperm nucleus GPx4 isoform. In contrast to the full GPx4 knockout, nGPx4−/− animals are viable and fully fertile, suggesting that the nGPx4 isoform is not responsible for the developmental defects observed when all the GPx4 isoforms are deleted (Imai et al, 2003; Yant et al, 2003). When spermatozoa from these nGPx4−/− animals were investigated more closely, they did not show any obvious phenotype. When spermatozoa from nGPx4−/− animals were compared to those of wild-type (WT) animals, it appeared that there was a delay in the completion of posttesticular sperm nucleus compaction. In the caput epididymis of nGPx4−/− animals, sperm nuclei were less compacted than spermatozoa from the caput epididymis of WT animals. This delayed compaction was resumed later on, because there was no difference in the state of sperm nuclei compaction for spermatozoa collected from the cauda compartment of nGPx4−/− and WT animals (Conrad et al, 2005). These data support the idea that nGPx4 acts as a thiol peroxidase on thiol-containing sperm nuclear protamines in the caput compartment of the epididymis. The fact that normal sperm-DNA compaction is recovered in spermatozoa stored in the cauda compartment of the nGPx4−/− animals suggests that one or more other thiol peroxidases most likely compensate for the lack of nGPx4 expression as the sperm cells travel along the epididymal tubule. Another possibility is that the cytosolic isoform (cGPx4), which is still expressed in testis of nGPx4−/− mice, may partly back up nGPx4 deficiency because it is small enough to enter the nuclear pore (Weis, 1998; Conrad et al, 2005). Finally, although H2O2 is commonly believed to be rather inefficient in mediating -S-S-bridging directly, one cannot exclude the idea that spontaneous disulfide bridging occurs during epididymal migration of sperm cells providing there is enough H2O2 in the epididymal lumen to sustain it.