Epididymal fluid is a gel-like and viscous dynamic phase compartment, which can be characterized as a giant membraneless organelle (Cornwall, 2009; Shum et al., 2011; Robaire and Hinton, 2015; Tecle and Gagneux, 2015; Gervasi and Visconti, 2017; Zhou et al., 2018; Cornwall et al., 2019). The intraluminal fluidic contents are compartmentalized distinctly according to anatomical regions of the epididymal tubule, as revealed by a glance at the region-specific patterns of different types of epithelial cells and their absorption and secretory machinery, as well as the changes of surface contents on the surface of epididymal spermatozoa (Turner et al., 2003; Ecroyd et al., 2004b; Johnston et al., 2005; Pastor-Soler et al., 2005; Yudin et al., 2005; Finger et al., 2006; Jelinsky et al., 2007; Turner et al., 2007; Shum et al., 2011; Tollner et al., 2012; Shum et al., 2014; Robaire and Hinton, 2015; Zhou et al., 2018). The mechanisms underlying the formation and maintenance of epididymal fluid compartmentalization are not fully understood yet. However, liquid-liquid phase separation, a widespread mechanism for cells to handle membraneless matrices (Hyman et al., 2014; Shin and Brangwynne, 2017) suggest that this process could play a role in regulation of epididymal luminal microenvironment.

The concentrations of all the epididymal fluid components reflect their dilution in water, whose volume in turn is adjusted as a function of the bathed ion concentrations; thus water transport is important for the optimal microenvironment in each compartment of the male excurrent duct, ensuring sperm maturation. And thus, water transport also influences Ca²⁺ homeostasis in the epididymal lumen, despite the regulation mechanism and the physiological implications still largely remain to explore. Proteins known to be relevant for water transport and electrolyte balance in epididymis, including cystic fibrosis transmembrane conductance regulator (CFTR) (Wong, 1998; Ruan et al., 2012), aquaporins (Stevens et al., 2000; Pastor-Soler et al., 2001; Cheung et al., 2003; Pastor-Soler et al., 2005; Da Silva et al., 2006), and the Na⁺-K⁺-pump and associated transporters (Breton et al., 1998; Bagnis et al., 2001; Wong P. Y. D. et al., 2002; Kujala et al., 2007; Zuo et al., 2011), have important role in maintaining epididymal function and male fertility (Wagenfeld et al., 2002; Xu et al., 2003; Yeung et al., 2004; Pruneda et al., 2007; Krausz and Riera-Escamilla, 2018; Xu et al., 2018).

Water is the medium of numerous hydrolysis processes in biological systems and carbon dioxide and its hydrolyzed derivative bicarbonate are not only the metabolic products but also an acid-base balancers in the body. In epididymis, bicarbonate-related acid-base balance is an essential process for acidification of luminal fluid, which is critical for maintaining spermatozoa in a quiescent state (Au and Wong, 1980; Carr and Acott, 1984; Breton et al., 1996; Newcombe et al., 2000; Pastor-Soler et al., 2005; Shum et al., 2011; Zuo et al., 2011). Hence, multitude of proteins must act in concert to establish and maintain optimal epididymal volume, electrolyte balance, and pH. The orchestration of these multiple proteins and their functions involves a complex network of cell-cell crosstalk involving autocrine, paracrine and lumicrine signaling (Kim et al., 2015; Robaire and Hinton, 2015; Gao da et al., 2016; Battistone et al., 2019; Ma et al., 2019; Kiyozumi et al., 2020). Figure 2 shows a basic model for the epithelial cell functions involved in epididymal fluid regulation, including water and ion transport, acid-base balance, and calcium homeostasis.

Table 1 summarises the concentrations of most inorganic ions contained in epididymal fluid, based on data published in the literature (Levine and Marsh, 1971, 1975; Levine and Kelly, 1978; Hinton and Setchell, 1980; Jenkins et al., 1980; Turner et al., 1984; Caflisch and DuBose, 1990; Hinton, 1990; Caflisch, 1992; Cooper et al., 1992; Clulow et al., 1994; Stoltenberg et al., 1996; Clulow et al., 1998; Sorensen et al., 1999; Newcombe et al., 2000). The high K⁺ ion concentration, up to approximately 55 mM in the distal cauda region (approximately ten-fold higher than in serum), and the low concentration of Ca²⁺ (down to 0.25 mM in the cauda region, approximately ten-fold lower than in serum) are notable. These special properties indicate that the microenvironment for sperm maturation is uniquely formed and maintained as spermatozoa transit the epididymal tubule.

In further characterization of the mechanisms regulating Ca²⁺ absorption in the epididymis, our research has demonstrated that the TRPV6-associated extracellular pH-sensitive Ca²⁺ conductance functions in an electrically coupled manner with the intracellular Ca²⁺-sensitive chloride channel TMEM16A in isolated single principal cells of the distal part of the rat epididymis (Gao da et al., 2016) (see Figure 2). The interplay between TRPV6 and TMEM16A does not only regulate Ca²⁺ resorption but also Ca²⁺-dependent anionic current, a driving force of fluid secretion (Wong P. Y. D. et al., 2002). Our research also found that TRPV6-like Ca²⁺ conductivity is regulated by negative membrane potential and is sensitive to extracellular pH and divalent ions (Ca²⁺ and Mg²⁺), all of these physiological features are uniquely formed and maintained in the epididymis. Therefore, the coupling of TRPV6 with TMEM16A makes this coupler a key player of epididymal secretion. TMEM16A has been described to mediate epithelial mucin secretion, including in response to inflammation (Huang et al., 2012; Lin et al., 2015; Zhang et al., 2015; Benedetto et al., 2019). Our research has shown that TRPV6 co-localizes with TMEM16A in the apical membrane of the rat epididymal epithelium, as well as in some epididymosomes of the very proximal and the very distal rat epididymis (Gao da et al., 2016). These findings raise the possibility of TRPV6-TMEM16A contributing to the secretion of epididymosomes from the epididymal cells, an epididymal cell-to-cell communication mechanism that is essential for sperm maturation as well as transgenerational paternal epigenetic information transfer (Sullivan et al., 2005; Lotvall et al., 2014; Belleannee, 2015; Chen et al., 2016a; Chen et al., 2016b; Sharma et al., 2016; Tkach and Thery, 2016; Sharma, 2019; Chan et al., 2020).

It is known that the epididymal luminal fluid is acidic and under resting physiological conditions, it maintains a pH range of about 6.4–6.8 in the corpus and cauda epididymidis of rodents (see Table 1). Our research has shown that in single cauda epithelial cells isolated from rat epididymis, the TRPV6-related conductance is inhibited at acidic pH (for example, pH 6.4) and activated at alkaline pH (for example, 7.4) (Gao da et al., 2016). In addition, the TRPV6 activity is also sensitive to the resting membrane potential, and its conductance is inwardly rectifying at negative membrane potentials and is negligible at positive potentials (Bodding, 2005; Hoenderop et al., 2005). Whereas the resting membrane potential of epididymal epithelial cells has been determined to range between –6 mV and –39 mV, with an average value of approximately –22 mV (Cheung et al., 1976; Gao da et al., 2016). Our research has shown that in these resting potential ranges, the Ca²⁺-permeable current in single cauda epithelial cells at extracellular pH 6.4, or even at pH 7.4 is indeed undetectable or negligible, respectively (Gao da et al., 2016). We speculate that under physiological conditions of acidic epididymal lumen, Ca²⁺ influx through TRPV6 channel alone is silent or at least negligible, while the electrical coupling between TRPV6 and TMEM16A promotes Ca²⁺ cation influx by accompanying Cl⁻ anion influx. However, the restricted localization of the extracellular pH- and Ca²⁺-dependent TRPV6-TMEM16A electrical coupler in the proximal and distal ends of the epididymis as well as in the epididymosomes not associated with extensive Ca²⁺ reabsorption could imply the presence of another calcium reabsorption mechanism or that TRPV6 and its coupler TMEM16A have a function other than Ca²⁺ reabsorption. Consistent with our speculation, a study of roosters found that those with epididymal stones had raised epididymal expression of TRPV6, although the connection between the two observations is still unclear (Oliveira et al., 2012).

To address whether in epididymidal regions where TRPV6-TMEM16A is absent, another Ca²⁺ reabsorption mechanism is active, our research has found vitamin K-dependent γ-glutamyl carboxylase (γ-carboxylase or GGCX)-mediated carboxylation of the substrate MGP to play a role in Ca²⁺ homeostasis in the epididymis (Ma et al., 2019). Consistent with our hypothesis, we found that GGCX-MGP expression was enriched in the epididymal regions with minimal TRPV6-TMEM16A coupler expression. Using a rat model of warfarin-induced vitamin K2 deficiency, we found that vitamin K2-dependent GGCX and its γ-carboxylation substrate MGP were essential for epididymal Ca²⁺ homeostasis, sperm maturation and thereby male fertility. Warfarin-induced vitamin K2-deficiency in rats leads to disruption of GGCX-mediated carboxylation of MGP, which leads to accumulation of stress granules and Ca²⁺ in the epididymal epithelium and lumen, and ultimately reduces male fertility (Ma et al., 2019).

Vitamin K is a known key modulator of Ca²⁺ homeostasis. By facilitating GGCX-mediated γ-carboxylation of matrix proteins, including MGP, it maintains bone health and prevents vascular ectopic calcification (Weber, 2001; Wallin et al., 2008; Boraldi et al., 2013). MGP is a 14-kDa (103 amino acids) secretory protein containing five Ca²⁺-binding γ-carboxyglutamic acid (Gla) residues. It was first described as an extrahepatic matrix protein in the extracellular matrix of bone and cartilage and later detected in a wide variety of other tissues such as lung, heart, kidney and arterial vessel walls (Price et al., 1983; Fraser and Price, 1988; Hale et al., 1988). MGP is synthesized locally, at least in the vascular tissues, where it counterbalances ectopic mineral deposition (Murshed et al., 2004). The essential role of MGP has been revealed in mice lacking MGP, which die within 6–8 weeks of birth from large blood vessels ruptures, owing to massive vascular calcification (Luo et al., 1997). Humans with mutations in both alleles of the MGP gene develop Keutel syndrome, an autosomal recessive disorder characterized by abnormal calcification of soft tissues and female miscarriages (Munroe et al., 1999; Khosroshahi et al., 2014). The vitamin K antagonist, warfarin through inhibition of the vitamin K epoxide reductase complex 1 (VKORC1) to prevent carboxylation and activation of MGP. The treatment of normal mice with warfarin has been associated with a rapid calcification of elastic lamellae in arteries and heart valves that is reminiscent of the MGP-null phenotype in mice (Price et al., 1998). The importance of vitamin K status for MGP has also been confirmed in healthy people and in patients with long-term warfarin treatment (Rennenberg et al., 2010; Cranenburg et al., 2012). Importantly, the understanding of vitamin K-dependent mechanisms for local calcification regulation may also provide information about luminal calcium homeostasis in the epididymis.

Like other Gla proteins, MGP is activated by GGCX, which is the only gamma-carboxylation enzyme in the cell and which has no relevant homology to any known enzyme families (Tie et al., 2016). The Ca²⁺ binding function of MGP requires two post-translational modifications: γ-glutamate carboxylation and serine phosphorylation (Schurgers et al., 2007; Schurgers et al., 2013; Houben et al., 2016). The exact mechanism through which serine phosphorylation in MGP facilitates Ca²⁺ binding is unclear, and may involve indirect mechanisms such as promoting extracellular secretion of MGP (Wajih et al., 2004). In contrast, γ-glutamyl carboxylation is precisely related to increased Ca²⁺ binding affinity of MGP (Hackeng et al., 2001; Schurgers et al., 2005).

Our research has showed that MGP expression is high in the epididymis when compared with other organs containing epithelial-lining such as kidney and liver (Ma et al., 2019). In addition, GGCX and MGP co-localized in the vesicular structures on the luminal surface of epithelial and sperm membranes, potentially indicating that carboxylation takes place in the epididymal luminal microenvironment. These results may dicates that the vitamin K-dependent carboxylation and the MGP-associated prevention of Ca²⁺-deposition can reciprocally regulate epididymal luminal secretions. Since the expression of GGCX and MGP and their co-localization on vesicular granules, sperm surface, and in the cytoplasmic droplets progressively increases from the proximal epididymis distally, our study suggests that both proteins likely arose from the epididymosomes derived from epididymal epithelial cells. In view of the important role of epididymosomes in the male reproduction and the health of offspring (Sullivan et al., 2005; Chen et al., 2016a; Chen et al., 2016b; Sharma et al., 2016; Conine et al., 2018, 2019; Sharma et al., 2018), whereas GGCX-dependent carboxylation of MGP plays a vital role in life events (e.g., early embryo development) (Schurgers et al., 2013; Khosroshahi et al., 2014), the understanding the role of carboxylated-MGP dependent luminal calcium homeostasis in epididymis may provide insights into male reproduction and the health of offspring.

In one of our previous studies (Ma et al., 2019), western blotting with anti-MGP antibody on rat epididymis and kidney always detected a major band at approximately 32-kDa that did not correspond to the expected ∼12-kDa molecular weight of MGP (Figure 3). Other publications also noted this band and interpreted it as a non-specific band due to the excessive protein lysis sample (Lomashvili et al., 2011). However, the intensity of this ∼32-kDa band was strikingly decreased when the tissue protein extracts were preincubated with the MGP peptide (Figure 3A). In addition, the intensity of the 32-kDa band and a few weaker bands at higher molecular weight became intensified when the divalent ion chelator EDTA was added to the protein extracts (Figure 3B), suggesting that these bands may represent Ca²⁺-dependent protein aggregates. Supporting this hypothesis, the 32-kDa band intensity from an epididymal cells line changed in a biphasic Ca²⁺-dependent manner. When the Ca²⁺ was added to the protein extract at sub-millimolar concentrations, the intensity of the Ca²⁺-bound MGP-positive protein complex increased. Maximum intensity of the MGP-positive protein complex was observed after addition of 0.1 mM Ca²⁺, suggesting that at this concentration protein aggregation is favored. When the Ca²⁺-concentration was further increased to the millimolar range, the band intensity was reduced, presumably because millimolar Ca²⁺ does not favor formation of MGP-containing protein aggregates (Figure 3C). This Ca²⁺-dependent, biphasic protein aggregation matches the Ca²⁺ levels that have been reported for the epididymal luminal microenvironment and range from ∼0.25 to ∼0.9 mM (Jenkins et al., 1980; Turner, 1991; Clulow et al., 1994; Turner, 2002). These results suggest that MGP-containing protein aggregates form in the presence of sub-millimolar Ca²⁺, but not at millimolar Ca²⁺ concentrations.

The formation of the MGP-binding protein complex can be interpreted from the biochemical nature of the MGP protein itself (Figure 4). The inactive form of uncarboxylated MGP (ucMGP) protein has an isoelectric point of about 8.8, so that in the epididymal microenvironment of around pH 6.4–6.8 (Levine and Marsh, 1971; Da Silva et al., 2007), the five γ-glutamate Glu residues of ucMGP endow the protein with five positive charges and thus hydrophilic. When these Glu sites are carboxylated by GGCX, the protein becomes its active Gla form of carboxylated MGP (cMGP), in which the positive charges are neutralized, so cMPG is electrically neutral and hydrophobic. However, the carboxylated sites have strong Ca²⁺ binding ability, which permit cMGP to bind five Ca²⁺ ions (Hackeng et al., 2001; Huang et al., 2003; Schurgers et al., 2005; Boraldi et al., 2013), and thus carrying ten positive charges, greatly enhancing its solubility. Therefore, in the acidic epididymal lumen, inactive ucMGP is soluble but lacks Ca²⁺ binding ability, whereas the active cMGP is insoluble but highly adherent to Ca²⁺, which confers the aggregating ability. In other words, Ca²⁺ ions could promote Ca²⁺-MGP scavenging function. Ca²⁺-MGP chelator function could be a perfect scavenger for the aggregation of either organic or inorganic extracellular matter or both, and regulated by the changing luminal microenvironment, including nutrients supply, acid-base balance and luminal Ca²⁺ levels, as well as other constituents in epididymal fluid.

The epididymal luminal microenvironment has a complex composition and contains components that can interact with MGP other than Ca²⁺, such as Mg²⁺ and phosphorus-containing electrolytes (see Table 1). EDTA chelates divalent cations and its affinity for Mg²⁺ is higher than for Ca²⁺. The increased intensity of the MGP-positive protein aggregate bands after the addition of exogeneous EDTA to protein lysates (see Figure 3B) can be explained by preferential chelation of Mg²⁺-ions and promotion of MGP interaction with Ca²⁺ and subsequent protein aggregate formation (Ma et al., 2019). This hypothesis is consistent with the finding that Mg²⁺ and Ca²⁺ competitively bind to MGP (Roy and Nishimoto, 2002). This also indicates that under the low Ca²⁺ concentrations of the epididymal lumen, activated MGP preferentially binds to Ca²⁺ and other organic matter, as illustrated in Figure 4. In the presence of high Ca²⁺, MGP has an affinity for inorganic phosphorous compounds (Roy and Nishimoto, 2002), but this would not occur in the epididymal lumen–the microenvironment for the maturation of spermatozoa whose membrane surfaces are enriched with phosphorates in lipids and proteins (Hamamah and Gatti, 1998; Jenkins et al., 1980). Under conditions with abnormally high Ca²⁺ (see Figure 4), such as impaired GGCX activity after warfarin treatment, Ca²⁺ precipitated granules result (Ma et al., 2019) and other pathological conditions may occur (Tan et al., 2012). A related phenomenon may be the clinical observations that patients undergoing continuous hemodialysis have a higher incidence of epididymal stones (Guvel et al., 2004; Bozzini et al., 2013), although the correlation and mechanism between these two phenomena are still unknown.

As illustrated in Figure 5A, GGCX requires the reduced vitamin KH2 (VKH2) as an obligatory cofactor and its carboxylating activity. Vitamin K generally undergoes a pathway called the vitamin K cycle, in which vitamin K is reduced to VKH2 by vitamin K epoxide reductase (VKOR or VKORC1); VKH2 is then re-oxidized during γ-glutamyl carboxylation by GGCX to vitamin K epoxide (VKO), which is subsequently converted to VKH2 by VKORC1 to complete the VK cycle (Stafford, 2005; Rishavy and Berkner, 2012; Tie and Stafford, 2016).

GGCX is a membrane protein with a size of approximately 87.5 kDa and has five transmembrane domains, with its N-terminus facing the cytoplasmic side and the C-terminus facing the luminal side, in addition to two intracellular loops and two luminal loops, as illustrated in Figure 5B (Tie et al., 2000; Tie et al., 2003; Rishavy and Berkner, 2012; Tie and Stafford, 2016). It is believed that the second intracellular loop contains the binding sequence for the inactive form of MGP (Pudota et al., 2001). On the luminal side, a disulfide bond is formed at position of Cys-99 and Cys-450, which is the essential site for the epoxidation and carboxylation of GGCX (Pudota et al., 2000). This disulfide bond is on the same side as the sequence residues for vitamin K deprotonation (Rishavy et al., 2006) and as the glutamate and pro-peptide binding sites (Mutucumarana et al., 2000; Mutucumarana et al., 2003). On the basis of the nature of this redox-sensitive binding, as well as the epoxidation and deprotonation nature of GGCX, the carboxylation and the other reactions in the vitamin-K cycle are all affected by redox status and acid-base balance (Rishavy and Berkner, 2012). Both of these two physiological factors are known to be essential for sperm maturation and male fertility (Au and Wong, 1980; Carr and Acott, 1984; Breton et al., 1996; Newcombe et al., 2000; Vernet et al., 2004; Pastor-Soler et al., 2005; Aitken and Curry, 2011; Shum et al., 2011; Robaire and Hinton, 2015; Aitken, 2016). It is suspected that the GGCX activity and redox- and pH-sensitive processes in the epididymis reciprocally influence each other.

In our previous study, we identified a single nucleotide polymorphism (SNP), rs699664, in the GGCX gene of infertile men with asthenozoospermia (Ma et al., 2019). This SNP replaces an arginine³²⁵ with glutamine³²⁵ (Arg325Gln) in the GGCX protein (Kinoshita et al., 2007; Rieder et al., 2007). The same SNP (Arg325Gln) has been reported to be associated with bone mineral density in aged females, and increased GGCX activity for MGP and bone matrix protein osteocalcin (Kinoshita et al., 2007). The SNP is located in the same second intracellular loop of GGCX, in close proximity to where the binding site of vitamin K-dependent proteins (i.e., MGP) is, as illustrated in Figure 5B. It has been reported that all the five Glu residues of MGP are converted into active Gla in one round of stepwise action of GGCX (Rishavy and Berkner, 2012). This process enables MGP to bind Ca²⁺ ions, Ca²⁺ crystals and bone morphogenetic proteins (Nadeem et al., 2004; Schurgers et al., 2013).

While further investigation is required, we speculate that the positively charged arginine³²⁵ residue of GGCX is responsible for stabilizing the negatively charged Glu residue of MGP during the stepwise carboxylation process, which is promoted in the acidic conditions of the epididymal lumen. Such charge-charge interaction stability is lost as a result of the SNP, in which the charged arginine³²⁵ residue is replaced with the neutral glutamine³²⁵, thereby changing the binding preference of MGP. Accordingly, the charge-charge interaction stability in the wild-type protein ensures sufficient time for complete carboxylation of each residue in the substrates and avoids their incomplete activation (Rishavy and Berkner, 2012). Supporting this interpretation is the finding that the rs699664 SNP in GGCX results in higher carboxylase enzymatic activity in aged females, and was associated with osteoporosis (Kinoshita et al., 2007). We speculate that the rs699664 SNP has similar effects in the epididymis of asthenozoospermic infertile males, i.e., increases early release of incompletely carboxylated substrate, resulting in under-carboxylated MGP, impaired Ca²⁺-binding, disturbed Ca²⁺-homeostasis, and ultimately to impaired sperm maturation and to male infertility.