The success of acrosomal exocytosis, a complex process with a variety of inter-related steps, relies on the coordinated interaction of participating signaling molecules. Since the acrosome reaction resembles Ca2+-regulated exocytosis in neurons, we investigated whether cognate neuronal binding partners of the multi-PDZ domain protein MUPP1, which recruits molecules that control the initial tethering and/or docking between the acrosomal vesicle and the plasma membrane, are also expressed in spermatozoa, and whether they contribute to the regulation of acrosomal secretion. We observed that CaMKIIα colocalizes with MUPP1 in the acrosomal region of epididymal spermatozoa where the kinase selectively binds to a region encompassing PDZ domains 10-11 of MUPP1. Furthermore, we found that pre-treating mouse spermatozoa with a CaMKII inhibitor that directly blocks the catalytic region of the kinase, as well as a competitive displacement of CaMKIIα from PDZ domains 10-11, led to a significant increase in spontaneous acrosomal exocytosis. Since Ca2+-calmodulin releases CaMKIIα from the PDZ scaffolding protein, MUPP1 represents a central signaling platform to dynamically regulate the assembly and disassembly of binding partners pertinent to acrosomal secretion, thereby precisely adjusting an increase in Ca2+ to synchronized fusion pore formation.
Ca2+-regulated exocytosis, a series of molecular events including vesicle recruitment, tethering, docking, priming, and the final vesicle fusion event itself, is adapted to deliver molecules with high reliability and with the exact moment-to-moment requirements of a cell (Chapman, 2008; Jahn, 2004; Malsam et al., 2008; Rettig and Neher, 2002; Rizo and Rosenmund, 2008; Sudhof and Rothman, 2009; Wojcik and Brose, 2007). The core fusion machinery consists of a super-family of conserved membrane-bridging proteins, called SNAREs (soluble N-ethylmaleimide-sensitive attachment protein receptor), which form a stable membrane-trafficking complex composed of a complementary set of vesicular synaptobrevin (v-SNARE), and plasma membrane-associated syntaxin and SNAP25 (t-SNAREs) (for reviews, see Jahn and Scheller, 2006; Lang and Jahn, 2008). At the subcellular level, the precision and adjustability of Ca2+-regulated exocytosis is brought about by a dual protein-assembly strategy of components involved in the fusion reaction sequence. First, this includes clustering of SNAREs and assisting regulatory proteins (Malsam et al., 2008) in liquid-ordered membrane microdomains, commonly termed `lipid rafts' (Lang, 2007) which have been implicated in the recruitment and regulation of signal transduction and membrane traffic processes (Ikonen and Simons, 1998). Second, components of the fusion machinery interact directly with each other (Mochida, 2000; Sudhof, 1995), and/or with specialized scaffolding proteins (Fejtova and Gundelfinger, 2006), thereby forming complex signaling frameworks below the plasma membrane (Schoch and Gundelfinger, 2006). Interestingly, a special mode of competitive protein-protein interaction between the Ca2+-sensor protein synaptotagmin (Koh and Bellen, 2003) and the SNARE-associated complexin, also called synaphin (Brose, 2008; McMahon et al., 1995), has recently been shown to clamp the fully assembled fusion machinery at an intermediate pre-fusion stage (Sudhof and Rothman, 2009), thereby preventing spontaneous SNARE-promoted membrane fusion (Weber et al., 1998).
A special form of Ca2+-regulated exocytosis is the sperm acrosome reaction, a mandatory event in the fertilization process, initiating two essential events: release of hydrolyzing enzymes, necessary for sperm to penetrate the egg glycoprotein matrix, the zona pellucida (ZP) (Wassarman and Litscher, 2008), and exposure of the inner acrosomal membrane, which thereby becomes accessible for the egg plasma membrane during the ultimate fusion event (Florman et al., 2008; Harper et al., 2008). Although acrosomal exocytosis in sperm cells differs from exocytotic events in other cellular systems, mainly because the acrosome is a huge and single secretory vesicle, certain remarkable parallels have been outlined between the acrosome reaction in sperm and vesicle fusion in neurons and neuroendocrine or exocrine cells (Mayorga et al., 2007; Tulsiani and Abou-Haila, 2004). First, zona-pellucida-induced secretion of the acrosomal contents is accomplished by an elevation in intracellular Ca2+ (Florman et al., 2008; Publicover et al., 2007), characterized by two sequential Ca2+ transients (Darszon et al., 2005). Furthermore, isoforms of the SNARE protein family, as well as assisting regulatory proteins that are also associated with lipid raft-derived signaling platforms (Boerke et al., 2008; Travis et al., 2001; Tsai et al., 2007), are expressed in mammalian spermatozoa (Gerst, 1999; Kierszenbaum, 2000; Mayorga et al., 2007; Tomes, 2007). Beside these parallels of acrosomal secretion and regulated exocytosis in other secretory systems, there are some unique features of the acrosome reaction, whose underlying molecular mechanisms are still unclear. One of these characteristics is the formation of hundreds of fusion points between the plasma membrane and the outer acrosomal membrane, ensuring the efficient delivery of the entire acrosomal contents (Barros et al., 1967; Michaut et al., 2000; Zanetti and Mayorga, 2009). Although there are currently no experimental data to explain this special feature of synchronous membrane pore formation during acrosomal secretion, an intriguing solution for this conundrum could be a scenario in which distinct pre-assembled SNARE complexes (De Blas et al., 2005; Roggero et al., 2007; Tomes et al., 2005; Zarelli et al., 2009) are mechanically coupled in a functional protein network between the outer acrosomal membrane and the adjacent plasma membrane. However, if mechanically coupled SNARE complexes are the basis of the observed synchronized fusion pore formation, a single `unintended' fusion event could be able to trigger a `domino' effect. This would initiate a `zipper-like' biochemical chain reaction, which would ultimately lead to an irreversible and total loss of the one and only secretory vesicle. Therefore, mechanisms might exist to prevent a spontaneous acrosome reaction in resting sperm.
We have recently observed that the multi-PDZ domain protein MUPP1, which contains 13 potential protein-binding motifs and is abundantly expressed in brain tissue, is also present in the acrosomal region of spermatozoa of different mammalian species (Heydecke et al., 2006). In these cells, MUPP1 is involved in recruiting molecules that control the initial tethering and/or docking of the acrosomal vesicle at specific sites of the plasma membrane (Ackermann et al., 2008). In vitro, MUPP1 interacts with distinct isoforms of the neuronal Ca2+-calmodulin kinase II (CaMKII) (Krapivinsky et al., 2004), a holoenzyme of several protein subunits (α, β, γ and δ), encoded by four closely related genes (Hudmon and Schulman, 2002b). In analogy to the fact that the CaMKII is functionally active in recruiting synaptic vesicles to the active zone of the presynaptic nerve terminal (Greengard et al., 1993; Leal-Ortiz et al., 2008), the MUPP1-controlled recruitment of the acrosomal vesicle to the plasma membrane (Ackermann et al., 2008) might be mediated by MUPP1-associated CaMKII (Krapivinsky et al., 2004). However, it is also conceivable that a CaMKII-catalyzed phosphorylation reaction is responsible for `freezing' the acrosomal vesicle in an intermediate pre-assembled fusion state (Wang, 2008), thereby preventing accidental spontaneous acrosomal secretion. Therefore, in the present study, we investigated whether the MUPP1 binding partners CaMKIIα and/or CaMKIIβ (Krapivinsky et al., 2004) are expressed in spermatozoa, and whether these kinases contribute to the regulation of acrosomal secretion. We show that a CaMKIIα-MUPP1 complex is present in the acrosomal region of spermatozoa, and that this complex serves a crucial role in clamping the acrosomal fusion machinery. The observed mechanism could be of pivotal importance to preserve the fertilization ability of spermatozoa.
Identification of CaMKIIα in rodent spermatozoa
To investigate whether CaMKII isozymes found to interact with MUPP1 in vitro (Krapivinsky et al., 2004), are expressed in rodent spermatozoa, subtype-specific antibodies recognizing specific sequences within the C-terminal association region of either the α- or the β-isoform (Hudmon and Schulman, 2002b) were used. Western blot analyses were performed with Rattus norvegicus (rat) and Mus musculus (mouse) brain cortex, cortex-, testis- and epididymis-derived protein fractions and whole-sperm lysates of both rodent species. The anti-CaMKIIα antibody detected two predominant immunoreactive bands in rat (Fig. 1A, left panel) and mouse (Fig. 1A, middle panel) testicular tissue and in control extracts from brain cortex. Although the 52 kDa band probably represents the predominant neuronal CaMKII isoform (Wayman et al., 2008; Yamauchi, 2005), the 65 kDa protein might represent an α-isozyme created by alternative splicing (Hudmon and Schulman, 2002a). In lysates of isolated epididymal spermatozoa of rat and mice, the major neuronal CaMKII isoform was present in both rodent species; in addition, a degradation product with a molecular mass of about 40 kDa was detectable in mouse spermatozoa (Fig. 1A, right panel).
The CaMKIIβ isoform showed a different expression pattern in rat and mouse tissues: In rat (Fig. 1B, left panel) and mouse (Fig. 1B, middle panel) brain cortex, a β-immunoreactive band with the predicted molecular mass of the prevalent neuronal β-isozyme of about 60 kDa (Yamauchi, 2005) was detectable. However, in testicular tissue of both rodent species, the antibody only recognized a smaller 45 kDa immunoreactive band (Fig. 1B, left and middle panel); in isolated rat and mouse sperm lysates, the β-isoform was not detectable (Fig. 1B; right panel).
Colocalization of CaMKIIα and MUPP1
To examine whether one of the two CaMKII isozymes found to associate with MUPP1 are detectable at the cellular level, double immunofluorescence staining was performed by incubating epididymal mouse and rat spermatozoa with anti-MUPP1 and anti-CaMKIIα or anti-CaMKIIβ antibodies (Fig. 2). Previous studies have shown that an antibody that recognizes both the CaMKIIβ and CaMKIIγ isoforms led to a distinct staining of the principal piece of the flagellum of mouse spermatozoa (Schlingmann et al., 2007); however, application of a subtype-specific anti-CaMKIIβ antibody did not reveal a specific β-immunolabeling, in neither mouse nor rat spermatozoa (data not shown), thus confirming the observations made by immunoblot analyses (Fig. 1B). In contrast to the lack of CaMKIIβ expression, experiments with the CaMKIIα-specific antibody verified the expression of the CaMKIIα isozyme in mouse spermatozoa detected by immunoblotting (Fig. 1A). Immunostaining was seen in sperm heads and was restricted to the hook-shaped acrosomal region (Fig. 2A). The same crescent-shaped acrosomal labeling was obtained with the anti-MUPP1 antibody (Fig. 2A). Overlay of the confocal microscopy pictures (orange-yellow) indicated that MUPP1 and CaMKIIα were colocalized within the acrosomal region of mouse spermatozoa (Fig. 2A).
To investigate whether CaMKIIα showed a comparable subcellular localization in rat sperm cells, isolated epididymal rat sperm were subjected to indirect immunostaining using the anti-CaMKIIα antibody that recognizes the association region of the kinase (Fig. 1A, left panel). However, no apparent labeling was observed (data not shown) by the latter antibody and by two other antibodies that were also raised against the C-terminal region of CaMKIIα (data not shown), suggesting that in rat spermatozoa, the association domain of CaMKIIα is not immunocytochemically accessible. To test this assumption, an antibody directed against residues surrounding the phosphorylated Thr286 in the regulatory domain of the kinase was chosen (Hudmon and Schulman, 2002a; Hunter and Schulman, 2005). Positive immunostaining would not only confirm expression of this kinase isoform in rat spermatozoa, but would also indicate whether CaMKIIα is already in an active conformation. Fig. 2B illustrates the fluorescence labeling pattern of rat sperm stained with the anti-CaMKIIα-Thr286-P antibody. A crescent-shaped acrosomal staining and a faint labeling of the flagellum restricted to the midpiece region of the sperm tail (Fig. 2B) was discerned.
In cultured neurons (Du et al., 2006) and in transfected human embryonic kidney 293 cells (Suzuki et al., 2008) CaMKIIα is associated with raft-derived detergent-resistant membrane fractions. Since MUPP1, which binds CaMKIIα in in vitro translated systems (Krapivinsky et al., 2004), is also concentrated in Triton-X-100-insoluble plasma membrane microdomains derived from brain and sperm preparations (Ackermann et al., 2008; Krapivinsky et al., 2004), we asked whether both proteins were colocalized to detergent-insoluble membrane clusters in sperm cells. Therefore, we extracted rat brain cortex tissue and freshly isolated epididymal rat spermatozoa with Triton X-100. Subsequently, equivalent amounts of total microsomal starting material (P50) and detergent-insoluble (PTx) and soluble fractions (STx) were separated by SDS-PAGE and probed by immunoblotting using the anti-CaMKIIα antibody generated against the C-terminal association region. Fig. 2C shows that in cortex- (left panel) and in sperm-derived preparations (right panel), total CaMKIIα as well as the raft marker protein caveolin-1 (Anderson, 1998) co-migrated with the Triton-X-100-insoluble pellet, indicating colocalization with MUPP1 in detergent-resistant membrane fractions (Ackermann et al., 2008). A similar lipid raft targeting was also observed for the sperm-derived Thr286-P CaMKIIα (data not shown).
Recent studies have shown that, during capacitation, lipid-raft-associated key proteins of the sperm exocytotic machinery e.g. SNARE proteins, accumulate at the apical tip of the sperm head, a location that reflects their functional role in acrosomal secretion (Gadella et al., 2008; Nixon et al., 2009). Since raft association upon capacitation might thus correlate with a physiological function of a molecule in acrosomal exocytosis, raft targeting of CaMKIIα was assessed in capacitated spermatozoa. Fig. 2D (left panel) shows that capacitation of rat sperm did not reduce association of CaMKIIα with the detergent-insoluble pellet. Correspondingly, the major portion of Thr286-P CaMKIIα was raft-associated in fully capacitated spermatozoa (Fig. 2D, right panel), although the non-raft membrane sample of capacitated sperm also showed minor immunoreactivity.
Epididymal spermatozoa already exhibit Thr286-P CaMKIIα
So far, our results illustrate that the α-isoform of CaMKII exhibits the same subcellular expression pattern in mouse and rat spermatozoa and additionally indicate that a Thr286-P and thus autonomously active conformation of CaMKIIα (Hudmon and Schulman, 2002b) already exists in non-capacitated epididymal sperm. The complex transformation process of sperm capacitation is characterized by numerous phosphorylation reactions, and an increase in cytosolic Ca2+ (Breitbart and Naor, 1999; Tulsiani et al., 2007; Witte and Schafer-Somi, 2007). Since both of these reactions are necessary for CaMKII activation (Hunter and Schulman, 2005; Yamauchi, 2005), the question arises as to whether the proportion of phosphorylated CaMKII increases during the process of capacitation. To address this question systematically, Thr286-P CaMKIIα was determined in rat epididymal spermatozoa isolated in HS buffer devoid of capacitation-inducing NaHCO3-BSA (Fig. 3A, left panel) as well as in HS buffer supplemented with NaHCO3 and BSA (Fig. 3A, right panel). In addition Ca2+-calmodulin (Ca2+/CaM)-inducible phosphorylation of CaMKIIα (Hudmon and Schulman, 2002b; Yang and Schulman, 1999) was assessed for non-capacitated and in vitro capacitated sperm. To test whether the total amount of CaMKIIα changed upon capacitation, the anti-CaMKIIα antibody, which was successfully applied for western blot analysis (Fig. 1A) was used (Fig. 3C). Representative immunoblots shown in Fig. 3A confirm that Thr286-P CaMKIIα is already present in freshly isolated uncapacitated rat spermatozoa (Fig. 3A, left panel). Furthermore, we observed that neither capacitation nor application of Ca2+-calmodulin caused a significant increase in the proportion of Thr286-P kinase in rat germ cells (Fig. 3A, right panel). To further support the notion that CaMKIIα is already in an active conformational state before capacitation, mouse spermatozoa were also examined for CaMKIIα phosphorylation. Freshly isolated epididymal mouse sperm also contained a phosphorylated and catalytically active state of CaMKIIα whose amount did not increase upon Ca2+-calmodulin incubation (data not shown), indicating that in epididymal uncapacitated spermatozoa of both rodent species, CaMKIIα is already in an autonomous activation state.
CaMKII inhibition increases spontaneous acrosomal secretion rates
Since the acrosomal location (Fig. 2A,B) and targeting of CaMKIIα to detergent-resistant membrane microdomains upon capacitation (Fig. 2D) suggest that CaMKIIα is involved in acrosomal exocytosis (Gadella et al., 2008; Nixon et al., 2009), initial experiments were created examining the effect of the anti-CaMKIIα antibody recognizing the C-terminal association domain on acrosomal secretion. Furthermore, the effect of commonly used membrane-permeable CaMKII inhibitors (KN62, KN93), which prevent Ca2+-calmodulin binding and thereby activation (Hudmon and Schulman, 2002a; Hunter and Schulman, 2005), was assessed. However, neither the anti-CaMKIIα antibody, nor the KN drugs KN62 or KN93 (data not shown) significantly affected the rate of acrosomal secretion. In addition, the CaMKII inhibitor autocamtide-2 inhibitory peptide II (AIPII) was used, because capacitated mouse sperm already contain a Thr286-P CaMKII holoenzyme (data not shown), so that KN drugs are unable to competitively interfere with the calmodulin-binding site, and the antibody-recognition region might be masked upon oligomerization of the association region during kinase activation (Rosenberg et al., 2006). AIPII mimics the autoinhibitory region within the regulatory region of CaMKII and thus competitively binds and inhibits the catalytic domain of CaMKII (Hudmon and Schulman, 2002a; Ishida et al., 1998; Malinow et al., 1989). Fig. 4A summarizes the quantified effect of different concentrations of AIPII (Gardner et al., 2007; Hojjati et al., 2007) on spontaneous acrosomal exocytosis and the rate of acrosomal secretion induced by the Ca2+ ionophore A23187. Maintenance of CaMKII in a catalytically inactive state did not affect Ca2+-induced acrosomal secretion. However, the spontaneous rate of acrosomal exocytosis was enhanced in a dose-dependent manner. A similar observation was made after pre-incubating capacitated mouse spermatozoa with different concentrations of the calmodulin antagonist W7 (Fan et al., 2003; Hidaka et al., 1981; Si and Olds-Clarke, 2000): Fig. 4B shows that W7 dose-dependently triggered spontaneous acrosomal secretion. However, in contrast to AIPII (Fig. 4A) the calmodulin antagonist also significantly enhanced acrosomal exocytosis induced by the Ca2+ ionophore, suggesting that calmodulin modulates more than one reaction event during the sequential process of acrosomal exocytosis (Mayorga et al., 2007).
Testicular CaMKIIα binds to PDZ domains 10-11 of MUPP1
In synapses of hippocampal neurons, CaMKII and the GTPase-activating protein SynGAP are brought in close proximity by direct physical interaction with the PDZ domains of MUPP1 (Krapivinsky et al., 2004). The observation of a colocalization of CaMKIIα and MUPP1 in the acrosomal region of rodent spermatozoa (Fig. 2A), together with the finding of co-migration of the two proteins in sperm-derived detergent-resistant membrane microdomains (Fig. 2C,D), raises the question of whether CaMKII and MUPP1 form stable protein complexes in testicular tissue. To address this, GST pull-down experiments were performed with lysates of testicular tissue or isolated spermatozoa and various GST-fused MUPP1 fragments that collectively represent the entire MUPP1-scaffolding protein (Lee et al., 2000) (Fig. 5A). Fig. 5B shows that in testicular tissue, the anti-CaMKIIα antibody, which was generated against subtype-specific sequences within the association domain of CaMKIIα, detected two α-immunoreactive bands, as described above (Fig. 1A). Examining the CaMKIIα-binding ability of MUPP1, we found that MUPP1 interacts with testicular CaMKIIα. Binding was only detectable with a GST-fusion construct comprising PDZ domains 10-11 of MUPP1 (GST M10-11) (Fig. 5B,D), whereas other MUPP1 PDZ-binding domains did not show any detectable interaction. Since CaMKII binds MUPP1 only in its Ca2+-calmodulin free state in vitro (Krapivinsky et al., 2004), half of the testicular preparation used for the GST pull-down in Fig. 5B was pre-treated with Ca2+-calmodulin, and subsequently, MUPP1-precipitated proteins were assayed for CaMKIIα binding. Fig. 5C documents that in the presence of Ca2+/calmodulin, GST M10-11 failed to pull down CaMKIIα, suggesting that Ca2+-occupied calmodulin dissociates CaMKII from MUPP1. To test whether CaMKIIα binding to MUPP1 depends on the activity of the kinase, half of the testicular tissue in Fig. 5D was pre-incubated with the CaMKII inhibitor AIPII. The results presented in Fig. 5E show that AIPII did not affect binding of CaMKIIα to GST M10-11, suggesting that recruitment of the kinase by MUPP1 does not depend on the catalytic activity of the kinase.
To examine whether CaMKIIα in spermatozoa was also associated with MUPP1, GST pull-down experiments were performed with isolated epididymal sperm cells, and subsequently, precipitates were probed for CaMKIIα binding. Fig. 5F shows that CaMKIIα also interacts with MUPP1 in isolated germ cells and that this interaction is also focused to GST M10-11.
To examine the functional significance of the observed MUPP1-CaMKIIα interaction for acrosomal secretion, the MUPP1-CaMKII complex was disrupted by pre-treating permeabilized spermatozoa with the purified MUPP1-M10-11 GST fusion protein, thereby competitively displacing CaMKIIα from GST M10-11 of MUPP1, and subsequently the spontaneous rate of acrosomal secretion was quantified. Since GST alone slightly but not significantly increases acrosomal secretion (Lopez et al., 2007) (data not shown), the specificity of CaMKIIα-MUPP1 GST M10-11 interaction was assessed in parallel approaches in which permeabilized sperm cells were pre-incubated with equal amounts of the MUPP1-M12-13 GST fusion protein. Fig. 6 shows that prevention of MUPP1-CaMKII interaction led to a comparable effect to that observed upon inhibition of CaMKII by AIPII (Fig. 4A) and blockage of calmodulin by W7 (Fig. 4B). Competitive displacement of CaMKIIα from endogenous MUPP1 led to a significant enhancement of spontaneous acrosome reaction (P<0.01), which was not detectable for the MUPP1-M12-13 GST fusion protein.
The acrosome reaction, a multistage reaction process that resembles Ca2+-regulated exocytosis in somatic cells, is specialized to maintain the reliable fusion of the large acrosomal vesicle by a zipper-like fusion pore formation spread out from the initial contact site with the egg glycoprotein matrix. However, the molecular mechanisms governing precisely timed, Ca2+-triggered synchronous membrane fusion, has only been partially defined. The results of the present study provide evidence that the CaMKIIα isozyme forms a protein complex with the multi-PDZ domain protein MUPP1 in the acrosomal region of rodent spermatozoa and that this complex is involved in preventing `accidental' agonist-independent acrosomal secretion.
The subcellular localization of CaMKIIα in the flagellum (Fig. 2B) and the head of spermatozoa (Fig. 2A,B) is consistent with a functional role of Ca2+-calmodulin kinase II in sperm motility and/or acrosomal secretion. The results of the present work demonstrate that autonomously active CaMKIIα negatively regulates acrosomal exocytosis in spermatozoa (Fig. 4). The finding that CaMKII prevents ZP-independent acrosomal exocytosis, and the fact that the kinase is already active in uncapacitated epididymal spermatozoa (Fig. 2B, Fig. 3) might explain the described relatively low rate of spontaneous loss of the acrosomal vesicle (Wassarman and Litscher, 2008). In presynaptic terminals, CaMKII is known to regulate synaptic transmission (Benfenati et al., 1993; Chi et al., 2003; Leal-Ortiz et al., 2008; Schiebler et al., 1986) by phosphorylating serine or threonine residues of target proteins, thereby modulating the function of these proteins (Wayman et al., 2008). An inhibitory effect of CaMKII on vesicle fusion is postulated to be mainly realized by the activation of large-conductance Ca2+-activated K+ channels [big potassium (BK) channels] (Wang, 2008), which represent negative key regulators of synaptic transmission (Liu et al., 2007). Although different types of K+ channels have been found to be expressed in mammalian spermatozoa (Darszon et al., 2006; Navarro et al., 2008), including BK channels encoded by the Slo gene family (Martinez-Lopez et al., 2009; Navarro et al., 2007; Schreiber et al., 1998), further experiments are needed to investigate whether K+ channels in germ cells are actually phosphorylated by CaMKII, and whether such covalent modification has any in vivo relevance for acrosomal exocytosis.
An indispensable prerequisite for ultimate membrane fusion is the release of the CaMKIIα-controlled clamp by an increase in free Ca2+. In this context, it is worth mentioning that Ca2+-calmodulin has been found to dissociate CaMKIIα from MUPP1 (Krapivinsky et al., 2004) (Fig. 5C), whereas inhibition of the catalytic activity of the kinase (Fig. 5E) does not induce kinase release (Fig. 7). In sperm, the Ca2+ controlled `unfreezing' of the MUPP1-CaMKII fusion clamp might either be triggered by the ZP3-induced initial fast and transient increase in Ca2+ that is mediated by voltage-gated Ca2+ channels (Arnoult et al., 1999), or by the second sustained efflux of Ca2+ from the acrosomal Ca2+ store, which is necessary to elicit ultimate fusion pore formation (De Blas et al., 2005; Herrick et al., 2005). Therefore, we suggest that CaMKIIα modulates proteins participating in distinct stages of the sequential Ca2+-controlled secretion pathway. On the one hand, it is conceivable that CaMKIIα inactivates proteins of intracellular transduction cascade/s stimulated upon ZP binding (Florman et al., 2008), thereby stabilizing the inactive cis-SNARE-complex in capacitated resting sperm (Zarelli et al., 2009). On the other hand, it is also feasible that target substrates of CaMKIIα are operative in triggering the final fusion step(s) of acrosomal exocytosis, such as the inositol-1,4,5-trisphosphate receptor (Aromolaran et al., 2007; Bare et al., 2005), or the t-SNARE syntaxin (Ohyama et al., 2002). However, the complexin- or synaptotagmin-mediated fusion clamp of the primed SNARE fusion machinery (Sudhof and Rothman, 2009) might also be a promising CaMKII-controlled `interlocking' target. Complexin is expressed in mammalian spermatozoa (Redecker et al., 2003; Roggero et al., 2007; Zhao et al., 2007); furthermore, synaptotagmin is known to present a target substrate for CaMKII phosphorylation (Abbott and Ducibella, 2001; Hilfiker et al., 1999). Thus, it will be important for future studies to investigate whether CaMKIIα phosphorylates and thus inactivates fusion components involved in Rab3A-elicited re-assembly of SNAREs leading to the primed membrane-bridging trans-SNARE complex (De Blas et al., 2005) (Zarelli et al., 2009), and/or whether target substrate(s) of CaMKII are involved in blocking ultimate fusion pore formation by either inhibiting the release of Ca2+ from the acrosome and/or by blocking the subsequent synaptotagmin-mediated release of the complexin interlock (Fig. 7).
An additional observation of the present work is that CaMKII is colocalized with MUPP1 in Triton-X-100-insoluble membrane fractions (Fig. 2C,D), and that direct interaction with a region consisting of the PDZ10-11 is necessary for the ability of CaMKII to prevent agonist-independent loss of the acrosomal vesicle (Figs 5, 6). Thus, MUPP1 might persistently organize activated CaMKII and possibly also its target substrate(s) in close proximity to a defined space between the outer acrosomal membrane and the plasma membrane (Fig. 7); a recruitment that might be further facilitated by lipid raft association (Fig. 2). The observation that lipid raft association and correct subcellular targeting are required for an effective and specific substrate phosphorylation of CaMKII has already been made (Strack et al., 1997; Tsui et al., 2005). Furthermore, binding to scaffolding proteins was found to prevent dephosphorylation and thus inactivation of CaMKII (Mullasseril et al., 2007). Although we do not know yet whether MUPP1 avoids kinase dephosphorlyation and whether CaMKIIα substrates also bind to MUPP1 (Fig. 7), the observation of a Ca2+-controlled assembly and disassembly of a MUPP1-CaMKIIα complex, in conjunction with the 13 binding motifs of MUPP1, offers an intriguing model of how MUPP1 and possible other interaction partner(s) (Fig. 7) might translate an increase in Ca2+ into final fusion pore building. This regulatory mode of action is of utmost importance for spermatozoa that only have a single opportunity to release their acrosomal contents. In this context, MUPP1 might be involved in passing a Ca2+-occupied calmodulin along the acrosomal reaction cascade. It has been suggested that calmodulin controls more than just a single step of the sequential membrane fusion process of the acrosome reaction, ranging from stimulation of T-type Ca2+ channels (Bendahmane et al., 2001; Lopez-Gonzalez et al., 2001) to inhibition of the final fusion event (Yunes et al., 2002). This hypothesis of multiple sites of action of calmodulin is extended by the present work demonstrating that the calmodulin antagonist W7 also increases spontaneous acrosomal secretion (Fig. 4B). We currently do not know whether Ca2+-occupied calmodulin mediates the observed `unlocking' of the fusion clamp by associating with either CaMKIIα or MUPP1 (Fig. 7). Nevertheless one might speculate that Ca2+-calmodulin affects several consecutive steps of the acrosome reaction, with MUPP1 guiding it from the initial recruitment and assembly of SNARE proteins, to the regulation of Ca2+ channels, to the fusion block and its subsequent release, up to final fusion pore formation. However, whether MUPP1 also interacts with itself by direct PDZ-PDZ binding (Harris and Lim, 2001), thereby forming a mechanically coupled scaffolding network below the plasma membrane, and whether such a mechanism is responsible for synchronized fusion pore building at hundreds of distinct loci in the sperm membrane (Barros et al., 1967) has to be explored in future studies.
Materials and Methods
Antibodies and reagents
Male adult mice and rats were raised at the Philipps-University, Marburg, or purchased from Charles River (Sulzfeld, Germany). To identify the α- and β-isoforms of CaMKII, as well as phosphorylated CaMKIIα, the following antibodies were used: for the CaMKIIα isoform, an anti-mouse IgG generated against the C-terminal region of rat CaMKIIα (residues 448-460) (BD Bioscience, Heidelberg, Germany) as well as an equivalent polyclonal (Santa Cruz, Biotechnology, Heidelberg) and a monoclonal anti-CaMKIIα IgG (Santa Cruz Biotechnology) were used. To identify the CaMKIIβ isozyme, a corresponding antibody also raised against the subunit-specific C-terminal association domain was used (Santa Cruz Biotechnology). For the detection of phosphorylated CaMKIIα, a mouse monoclonal antibody generated against a peptide surrounding the Thr286-P of rat origin (Santa Cruz Biotechnology) and a rabbit polyclonal antibody matching Thr286-P of mouse, rat and human CaMKIIα (Abcam, Cambridge, UK) (Pezet et al., 2008) were used.
The sources of other applied antibodies and chemicals are as follows: anti-caveolin1 IgG and donkey anti-goat horseradish peroxidase (HRP)-labeled IgG were obtained from Santa Cruz Biotechnology. HRP-conjugated goat anti-rabbit IgG was obtained from Bio-Rad (Munich, Germany); HRP-conjugated goat anti-mouse IgG was purchased from Amersham-Pharmacia (Freiburg, Germany). A biotinylated anti-mouse IgG as well as fluorescein avidin were purchased from Vector Laboratories (Burlingame, CA); Alexa-Fluor-546-conjugated goat anti-rabbit antibody was from Invitrogen (Karlsruhe, Germany). Anti-MUPP1 antibody was generated against a protein region between PDZ domains 5 and 6 of MUPP1 (Ackermann et al., 2008). The KN-CaMKII inhibitors 2-[N-(2-hydroxyethyl)]-N-(4-methoxy-benzenesulfonyl)amino-N-(4-chlorocinnamyl)-N-methylbenzylamine (KN93), as well as its inactive analogue 2-[N-(4-methoxybenzenesulfonyl)]amino-N-(4-chlorocinnamyl)-N-methylbenzylamine phosphate (KN92) (Marley and Thomson, 1996), and 1-[N,O-bis-(5-Isoquinolinesulfonyl)-N-methyl-L-tyrosyl]-4-phenylpiperazine (KN62) (Tokumitsu et al., 1990), which competitively interact with the calmodulin-binding site of CaMKII (Sumi et al., 1991) were from Calbiochem-Novabiochem (Bad Soden, Germany); likewise, the non-cell-permeable non-phosphorylatable CaMKII-peptide-substrate inhibitor autocamtide-2 inhibitory peptide II (AIPII) was derived from the autoinhibitory region of CaMKII (Ishida et al., 1995; Ishida et al., 1998) and its cell-permeable analogue fused to the antennapedia transport peptide sequence (Passafaro et al., 1999), as well as the CaM antagonist W7 [8 (N-86-aminohexyl)-5-chloro-1-naphthalenesulfonamide] (Hidaka et al., 1981), the phosphatase inhibitor okadaic acid (Cohen et al., 1990) and a protease inhibitor cocktail III. Recombinant calmodulin and ATP were from BioLabs (P6060S) (Hohenwestedt, Germany); the chemiluminescent substrates West Femto and West Pico ECL were from Thermo Scientific (Schwerte, Germany). The glutathione S-transferase (GST) fusion protein MUPP1-expression plasmids (pGEX-2T or pGEX-4T-1) presenting distinct MUPP1-PDZ domains [M1-3 (residues 118-504), M4-5 (residues 489-785), M6-9 (residues 780-1611); M10-11 (residues 1606-1830), M12-13 (residues 1825 to 2054)] were a gift from Ronald T. Javier (Department of Molecular Virology and Microbiology, Baylor College of Medicine, Houston, TX), and have been described elsewhere (Barritt et al., 2000; Lee et al., 2000). The pGEX-2T plasmid expressing GST alone was provided by Michael Bölker (Department of Genetics, Faculty of Biology, Philipps-University, Marburg, Germany). All other reagents, unless specified otherwise, were either purchased from Sigma (Deisenhofen, Germany) or Carl Roth (Karlsruhe, Germany).
Sperm preparation, capacitation and immunocytochemistry
Adult mouse and rat spermatozoa were isolated as described previously (Heydecke et al., 2006). To obtain spermatozoa of different capacitation status, motile sperm of cauda epididymis were collected either in HS working solution (30 mM HEPES, 135 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM glucose, 10 mM lactic acid and 1 mM pyruvic acid, adjusted to pH 7.4 with NaOH) or in capacitating buffer (HS, 0.5% BSA, 15 mM NaHCO3) and incubated for 90 minutes at 37°C in 5% CO2 (Ackermann et al., 2008). Indirect immunocytochemical approaches were performed as described (Fehr et al., 2007); control samples were incubated with 10% FCS in PBS. Double-fluorescence immunocytochemical staining of MUPP1 and CaMKIIα was performed by incubating samples of mouse spermatozoa for 48 hours with the mouse anti-CaMKIIα antibody recognizing the C-terminal association region and the rabbit anti-MUPP1 antibody (Ackermann et al., 2008). Subsequently excess of antibody was removed by three washes with PBS, and samples were incubated for 1 hour at room temperature with the biotinylated anti-mouse IgG. After three additional washes, a mixture of the Fluorescein-crosslinked Avidin and the Alexa-Fluor-546-conjugated anti-rabbit IgG was applied (1 hour at room temperature). After three additional washes in PBS, samples were coated with fluorescent mounting medium (DAKO, Cytomation, Hamburg, Germany) and examined under a Zeiss LSM 510 META laser scanning confocal microscope. The results of the double-staining experiments showed staining patterns and staining intensities similar to those obtained in performed single staining experiments.
Recombinant protein expression, purification and GST pull-down assay
GST fusion proteins (GST, GST-M1-3, GST-M4-5, GST-M6-9, GST-M10-11, GST-M12-13) (Lee et al., 2000) were obtained from isopropyl 1-thio-β-D-galactopyranoside (IPTG)-induced bacterial cultures and purified through affinity chromatography using glutathione-Sepharose 4B beads (Amersham-Pharmacia, Freiburg, Germany) according to the manufacturer's instructions. Purity of the eluates was verified by staining aliquot samples by Coomassie Brilliant Blue dye stained SDS-PAGE; protein concentration of pooled fractions was determined by Bradford (Bradford, 1976).
For GST pull-down assays, MUPP1-GST fusion-protein-coated glutathione-Sepharose 4B beads were used as matrix for binding of CaMKIIα prepared from rat testis or isolated spermatozoa, respectively. To this aim, each of the five purified MUPP1-GST fusion proteins (75 μg) covering the whole MUPP1 protein or GST alone were incubated overnight at 4°C with PBS-equilibrated glutathione-Sepharose beads (60 μl). Thereafter, the coated GST-glutathione beads were washed with PBS, and subsequently, the unspecific binding sites were blocked by incubating the beads for 1 hour at 4°C with 10% BSA in PBS. At the same time, tissue preparations were either sonicated in homogenization buffer containing 10 mM Tris-HCl, 3 mM MgCl2, 2 mM EDTA, pH 7.4, protease inhibitor cocktail (testis) or in 50 mM Tris-HCl, 150 mM NaCl, 2 mM EGTA, 0.5% NP40, pH 8.0, protease inhibitor cocktail, 0.1 mM phenylmethanesulfonyl fluoride solution (sperm), solubilized by stirring (30 minutes at 4°C) and centrifuged for 10 minutes at 1000 g to separate cell nuclei and intact organelles. The supernatants (S1) were subsequently cleaned from cell debris (10,000 g, 10 minutes, S2) and incubated overnight at 4°C under constant agitation on a rotary shaker together with fusion protein-loaded Sepharose beads. After extensive washing of the beads with PBS, samples were boiled in loading buffer (Laemmli, 1970) and subjected to SDS-PAGE and immunoblotting. To analyze the effect of Ca2+ and calmodulin on MUPP1-CaMKIIα interaction, one half of a freshly prepared S2 fraction of testicular tissue was incubated for 10 minutes at 30°C with 2 mM CaCl2, 1.2 μM calmodulin and 100 μM ATP (Yang and Schulman, 1999). To examine whether the CaMKIIα-MUPP1 interaction was activation dependent, one half of the testis tissue preparation was homogenized and solubilized in buffer supplemented with 50 μM non-cell permeable CaMKIIα inhibitor AIPII.
Acrosome reaction in intact and permeabilized spermatozoa
To induce acrosome reaction in unpermeabilized capacitated mouse spermatozoa, the Ca2+ ionophore A23187 was used to raise the intracellular Ca2+ concentration, as described recently (Heydecke et al., 2006). To elicit acrosome reaction in streptolysin-O (SLO)-permeabilized cells, the extracellular Ca2+ concentration was increased as described (Ackermann et al., 2008). Acrosome-reacted sperm were determined using Coomassie Brilliant Blue staining (Zeginiadou et al., 2000); for each experiment, at least 100 spermatozoa of each slide with blinded labels were scored using a Zeiss microscope (Axiovert 200M) or an Olympus microscope (CKX 31) equipped with bright-field light optics. Decoded samples were subsequently identified, and the percentage of spermatozoa with intact acrosome was calculated. To determine the functional role of CaMKII in the acrosome reaction, the effect of anti-CaMKIIα antibodies, a GST fusion protein containing the PDZ domains 10 and 11 of MUPP1 (GST M10-11, residues 1606-1830), one containing PDZ domains 12 and 13 (GST M12-13, residues 1825-2054) or GST alone, as well as a battery of cell-permeable protein kinase antagonists (KN93/KN92, KN62, AIPII, W7) were used. Stock solutions (10 mM) of KN92, KN93 and KN62 were prepared in DMSO and kept in the dark at –20°C until use; for each experiment, a control sample was included which was supplemented with DMSO concentrations equivalent to the maximum concentration of the applied inhibitor; final concentrations of DMSO never exceeded 0.01%. Non-membrane-permeable compounds or the anti-CaMKII antibodies were directly added after sperm permeabilization. After incubation for 20 minutes at 37°C, the acrosome reaction was induced and thereafter stopped by washing the germ cells. Spermatozoa were then fixed and their acrosomal status evaluated as described above. The spontaneous loss of acrosome was assessed by scoring the acrosomal status of sperm samples exposed only to the DMSO or SLO solution. In experiments in which sperm were pre-treated with SLO, experiments were included only when the rate of acrosome reaction induced by the addition of extracellular Ca2+ was higher than 5%, thus reflecting a successful permeabilization.
Values are presented as means ± s.e.m. of independent experiments with isolated sperm from different animals. Statistical analyses were performed using a paired Student's t-test. Levels of statistical significance were indicated by asterisks: *P<0.05; **P<0.01; ***P<0.001.
We thank Marga Losekamp for excellent technical assistance and Hennig Stieve for critical reading of the manuscript. In addition, the authors wish to thank Dorke Meyer for her comments on preliminary versions of this manuscript, and Ronald T. Janvier for kindly providing the GST-PDZ constructs. This work was supported in part by the Hertie-Exzellenzprogramm Neurowissenschaften and the Universitätsklinikum Giessen/Marburg (Förderung gem. 2 Abs. 3 des Kooperationsvertrages). The experiments comply with Principles of Animal Care of the National Institutes of Health and with the current laws of Germany.