Introduction

Meiosis is a unique process consisting of molecular events aiming to ensure the successful segregation of genetic material into the daughter cells. Meiotic recombination is initiated during the leptotene stage by the formation of double-strand DNA breaks, which precede the initiation of homologous chromosome pairing and synapsis (Mahadevaiah et al. 2001). By the end of the next stage, zygotene, the process of synapsis is completed, followed by meiotic recombination at pachytene on fully synapsed autosomes. A protein structure termed the synaptonemal complex (SC) is associated with the axes of homologous chromosomes before and after their synapsis (Cohen et al. 2006).

Unlike autosomes, X and Y chromosomes are aligned only along a small pseudoautosomal region (PAR) of sequence homology, and undergo the process of meiotic sex chromosome inactivation (MSCI). XY chromatin is remodeled in a highly condensed domain known as the sex body, which becomes obvious at the nuclear periphery of pachytene spermatocytes (Hoyer-Fender 2003; Handel 2004; Turner 2007). The role of the sex body is still under debate, but its formation is thought to be important for the successful completion of meiotic progression and the production of normal, genetically balanced gametes (Martin 2008). Failure of MSCI has been suggested to be sufficient to trigger midpachytene apoptosis in several mouse mutant models (Mahadevaiah et al. 2008). Numerous proteins have been localized to the sex body, but a crucial role in MSCI has been recently reported for BRCA1, ATR kinase, and histone γH2AX (H2AX, a phosphorylated subtype of the H2A histones; Turner et al. 2004, 2005). Two different waves of H2AX phosphorylation have been described. The first one, directed by ATM kinase at leptotene, occurs in the whole nucleus and is the result of double-strand break (DSB) formation preceding meiotic recombination; the second one (at zygotene) takes place on the sex chromosomes and is under the control of ATR kinase. The latter event is characterized by the gradual disappearance of γH2AX from the entire nucleus, followed by its accumulation on the sex chromosomes. This process is coincident with the initiation of MSCI (Mahadevaiah et al. 2001; Fernandez-Capetillo et al. 2003). Thus, the γH2AX localization pattern can serve as a marker for substaging of early meiotic events and the timing of MSCI. Experimental data obtained in BRACA-1 null animals suggest that BRCA1 can be placed upstream of ATR kinase in the sequence of MSCI events, and is responsible for directing ATR to the sex chromosomes. In addition to the sex body, γH2AX, ATR and BRCA-1 also localize to any autosomal regions that fail to synapse, and are involved in the general mechanism of meiotic silencing of unsynapsed chromatin (MSUC) during both male and female meiosis (Turner et al. 2004, 2005; Baarends et al. 2005).

Although the importance of BRCA-1, ATR, and γH2AX in sex body formation has been demonstrated, the exact mechanism underling XY inactivation, as well as the role of other proteins in this process remains to be determined. It is unknown, for example, what determines the concentration of BRCA-1 on the sex chromatin or how BRCA-1 induces a new wave of H2AX phosphorylation. Recent data also suggest that, by some unknown mechanism, normal DSB formation is required for sex body formation, suggesting the regulation of the latter process on multiple levels (Bellani et al. 2005).

The progression of meiotic events and sex body formation are tightly regulated by various posttranslational modifications. Sumoylation (covalent modification by small ubiquitin-like modifiers, or SUMO proteins) has been implicated in a range of cellular pathways, such as protein–protein interactions, transcription regulation (repression in most cases), nuclear–cytoplasmic transport, and formation of specific heterochromatic nuclear subdomains (Gill 2004; Johnson 2004). There are four different members of the SUMO superfamily: SUMO-1, -2, -3, and -4. SUMO-1 shares only 45% homology with SUMO-2 and -3, which are 96% identical (often termed as SUMO-2/3). SUMO-4 has been identified as a kidney-specific protein (Schwartz and Hochstrasser 2003; Seeler and Dejean 2003; Bohren et al. 2004; Muller et al. 2004). SUMO-1 has recently been shown to localize to the sex body area of mouse spermatocytes, where it was first seen at the zygotene stage, suggesting a role for SUMO-1 in the process of MSCI (Vigodner and Morris 2005). The localization pattern of SUMO-2/3 has recently been shown to be similar to that of SUMO-1 in pachytene and diplotene spermatocytes; however, such data are missing for zygotene spermatocytes, at the time when MSCI is initiated (La Salle et al. 2008). The role of different SUMO isoforms during MSCI is still unknown.

In order to better understand the chain of molecular events and the role of sumoylation in MSCI, in this study, we have followed dynamic changes in γH2AX and SUMO localization patterns during early meiotic prophase in relation to the timing of MSCI. Confocal laser scanning microscopy (CLSM) was used as an analytical tool for visualizing numerous spermatocytes from the same development stage and for consecutively following meiotic progression inside the seminiferous tubules. We were able to demonstrate a very early appearance of SUMO-1 that preceded γH2AX accumulation over the sex chromosomes during their inactivation. In contrast to SUMO-1, SUMO-2/3 was undetectable in zygotene spermatocytes, suggesting a specific role for SUMO-1 during MSCI.

Materials and methods

Animals, reagents, and antibodies

Adult male C57Bl6 mice were obtained from Jackson Laboratory (Bar Harbor, ME, USA). Animals were housed in standard lighting (12 h light, 12 h dark) and allowed food and water ad libitum. They were maintained in facilities approved by the American Association for the Accreditation of Laboratory Animal Care. Procedures involving the use of animals strictly followed the Guidelines for Care and Use of Laboratory Animals set forth by the NIH.

An affinity-purified rabbit polyclonal antibody raised against full-length SUMO-1 protein was purchased from Affinity BioReagents (Golden, CO, USA; SUMO-1ABR antibody). A rabbit polyclonal antibody raised against the C-terminus of human SUMO-2/3 protein was purchased from Abcam (Cambridge, MA, USA). A monoclonal antibody against phosphorylated histone H2A.X (Ser139, γH2A.X) was purchased from Upstate Cell Signaling Solutions (Lake Placid, NY, USA). All remaining reagents were purchased from Sigma (St. Louis, MO, USA) unless otherwise noted.

Confocal microscopy of seminiferous tubules

Small portions of paraformaldehyde-fixed tubules were covered with Igepal (0.3%) in microfuge tubes (10 min), washed twice without centrifugation, and blocked in Image-IT™ FX Signal Enhancer (Molecular Probes, Eugene, OR, USA). Tubules were incubated with both SUMO-1 and γH2A.X antibodies overnight (4°C) at final dilutions of 1:100 and 1:300 in PBS/1% BSA for SUMO and γH2A.X, respectively. Tubules were rinsed in PBS, and then incubated with Alexa Fluor 488-conjugated goat anti-rabbit IgG and Alexa Fluor 633 goat anti-mouse IgG. After three washes, tubules were incubated overnight with propidium iodide (PI, 50 μg/ml). Using a pipette with a widened tip, a few tubules were placed into optical glass-bottomed dishes (MatTek Corporation, Ashland, MA, USA) and subsequently examined by Zeiss LSM 510 inverted confocal microscopy, using FITC and CY-3 filter sets. Confocal microscopy analysis was performed as we have reported previously (Vigodner et al. 2002; Vigodner and Morris 2005).

Longitudinal optical sections were produced at the layers of differentiating germ cells in seminiferous tubules and stage-specific changes were followed along the tubule.

The specific spermatogenic and meiotic stage was determined according to the morphology of germ cells (mostly spermatids), the ratio of spermatogonia to Sertoli cells, and the presence of meiotic figures in each particular tubule fragment, as previously described (Russell et al. 1990; Vigodner et al. 2002; Vigodner and Morris 2005).

Results

To more accurately determine and to compare the timing of SUMO-1and γH2AX appearance on the sex chromosomes, whole-tubule immunostaining with both anti-SUMO-1 and anti-γH2AX antibodies was followed by confocal microscopy analyses. CLSM on longitudinal optical tubule sections enabled focusing on numerous spermatocytes at the same development stage, thereby allowing their precise morphological characterization. In mouse spermatogenesis, 12 stages have been described (Russell et al. 1990). Premeiotic stages were characterized by the presence of both round and elongating spermatids in the deeper layers of the same tubule fragment (Fig. 1A). Preleptotene spermatocytes (PL) were identified by staging using the presence of step 8–9 elongating spermatids, mature testicular sperm being released into the tubule lumen (not shown), and maximal germ-to-Sertoli cell ratio, all characteristics of stages VIII–IX in mouse spermatogenesis (Vigodner et al. 2002; Vigodner and Morris 2005; Fig. 1B). Maturation of leptotene spermatocytes (Le) was identified based on the distinct morphology of elongating spermatids from step 8–9 and 10–11 present in the deep tubule layers (Fig. 1C and D, respectively). Zygotene spermatocytes (Zyg) were characterized on the basis of their correspondence to stage XII of mouse spermatogenesis. The stage was determined according to the presence of meiotic figures (Meio), and the formation of secondary spermatocytes and round spermatids in the deeper tubule layers (Fig. 1E). Pachytene spermatocytes showed a defined sex body area in the nuclear periphery (Fig. 1E4 (insert); Vigodner and Morris 2005).

Fig. 1
figure 1

Identification of different types of spermatocytes in mouse seminiferous tubules using confocal microscopy. Confocal microscopy was performed on laser-produced longitudinal optical sections obtained at different depths in the tubules; the results of immunostaining with the anti-SUMO-1ABR (green) and anti-γH2AX (blue) antibodies are illustrated. Cell nuclei are visualized by propidium iodide (red). (a) A premeiotic stage with spermatogonia (Sg) is shown. Pachytene spermatocytes (PS), round spermatids (RS) and elongating spermatids from step 14–15 (E sp st 14–15), characterizing this stage, are situated deeper in the tubules and partially seen in this focal plane. (b). Preleptotene spermatocytes (PL) were identified by staging using the presence of testicular sperm adjacent to and within the lumen, and morphology of elongating spermatids (not shown), as well as maximal germ-to-Sertoli cell ratio (Vigodner et al. 2002; Vigodner and Morris 2005). Se, Sertoli cell. (c) Early leptotene spermatocytes (Le) were identified according to the distinct morphology of step 8–9 elongating spermatids present in the deeper layers of the same tubule fragment. (d) Late leptotene spermatocytes (Le) undergoing the transition to zygotene (Ze) were identified according to the distinct morphology of step 10–11 elongating spermatids present in the deeper layers of the same tubule fragment. (E1–E4) Zygotene spermatocytes (Zyg) are shown. (E2) and (E4) are the identical fields matched with (E1) and (E3), respectively, in which green fluorescence alone is shown. The spermatocytes were characterized on the basis of their correspondence to stage XII of mouse spermatogenesis. The stage was identified according to the presence of meiotic figures (Meio), the formation of secondary spermatocytes and round spermatids in the deeper tubule layers and the presence of step 12 elongating spermatids. Arrows indicate SUMO-1-positive nuclear foci as they became apparent in zygotene spermatocytes. (E4 insert) Pachytene spermatocytes showed a defined sex body area in the nuclear periphery. Scale bar represents 10 µm

Full size image

Figure 2 summarizes the localization pattern for γH2AX and SUMO-1 in different types of spermatocytes. The cells presented were enlarged using the stage-specific images shown in Fig. 1 or similar images obtained at the same stages. Using anti-SUMO-1 and anti-γH2AX immunostaining, a specific γH2AX signal without any SUMO-1 signal was detected in the nucleus of spermatogonia (Sg) and preleptotene (PL) and leptotene spermatocytes (L) (Fig. 1A–D and 2A, B, C) At the same spermatogenic stages, however, SUMO-1 was localized to the sex body of PS, as previously described (PS, Fig. 1A, C, D; Vigodner and Morris 2005). The presence of γH2AX in spermatogonia suggests an initiation of protein phosphorylation during premeiotic stages, a finding that should be further investigated. The localization of γH2AX to the nucleus of early spermatocytes is consistent with its role in early recombination events. Interestingly, the γH2AX signal was low in preleptotene spermatocytes and intensified as spermatocytes progressed from stages VIII–IX (Fig. 1C and 2C, Leptotene I) to stages X–XI (Fig. 1D and 2D, Leptotoene II), a finding that suggests that the later stages are characterized by massive DNA double-strand break formation.

Fig. 2
figure 2

Dynamic confocal microscopic localization of SUMO-1 and γH2AX during initiation of meiotic prophase and MSCI. The cells presented were enlarged using the stage-specific images shown in Fig. 1 or similar images obtained at the same stages. For each type of spermatocyte, both the SUMO-1, γH2AX and PI superimposed image and SUMO-1 staining alone are shown. Using anti-SUMO-1 and anti-γH2AX immunostaining, a specific γH2AX signal without any SUMO-1 signal was detected in the nucleus of spermatogonia (panel A), preleptotene (panel B), early and late leptotoene spermatocytes (panels C and D). At the same spermatogenic stages, SUMO-1 was localized to the sex body of PS, as previously described (A, C, D; Vigodner and Morris 2005). Some stage XII spermatocytes showed abundant expression of γH2AX but no SUMO-1 expression (panel E, Zygotene I). As the stage progressed, SUMO-1-positive nuclear foci became apparent, while the γH2AX signal was still detected in the entire nucleus (panel F, Zygotene II). The SUMO-1 signal gradually intensified over the sex chromosomes, forming the so-called ‘tadpole-shaped’ structures that characterize the sex chromosomes at zygotene (panel G, Zygotene III). This stage was followed by the concentration of γH2AX in the SUMO-1-positive area and the gradual disappearance of the protein from the rest of the nucleus (panel H, Zygotene IV). In pachytene spermatocytes, γH2AX and SUMO-1 were found to be mostly co-localized over the formed sex body area, as was described previously (panel I; Vigodner and Morris 2005). Scale bar represents 10 µm

Full size image

Based on SUMO-1 and γH2AX localization patterns, it was possible to further subdivide spermatocytes at stage XII of spermatogenesis into several sub-stages. Some spermatocytes at this stage still showed abundant γH2AX expression but no SUMO-1 in their nucleus (Fig. 2E, Zygotene I). As the stage progressed, SUMO-1 foci became apparent, while the γH2AX-positive signal was still detectable in the entire nucleus, with the γH2AX pattern corresponding to the time when X and Y chromosomes are still active (Turner 2007; Fig. 1E1–E4, Zyg, arrows; Fig. 2F, Zygotene II). The SUMO-1 signal gradually intensified over the sex chromosomes, forming the so-called ‘tadpole-shaped’ structures that characterize the sex chromosomes at zygotene (Fig. 2G, Zygotene III). This stage was followed by the concentration of γH2AX in the SUMO-1-positive area and the gradual disappearance of the protein from the rest of the nucleus. This time was shown to be coincident with MSCI (Fig. 2H, Zygotene IV).

In pachytene spermatocytes, γH2AX and SUMO-1 were observed mostly co-localized over the formed sex body area, as was described previously (Fig. 1E4, insert; Fig. 2I, Pachytene; Vigodner and Morris 2005).

Taken together, our results indicate that SUMO-1 appearance on the sex chromosomes precedes accumulation of γH2AX in the same nuclear area, and therefore occurs either before or very early during MSCI.

Interestingly, using an antibody against SUMO-2/3, we were unable to detect the protein in zygotene spermatocytes, a finding that differed from that obtained for SUMO-1 (Fig. 3A, A1, B, B1). As with SUMO-1, however, early, mid-, and late pachytene spermatocytes did show SUMO-2/3, seen mostly within the sex body area. Some centromeric heterochomatin was also stained, a pattern that was more evident using isolated cell bioimaging and is consistent with recent reports on pachytene spermatocytes (Fig. 3C and insert; La Salle et al. 2008). Using confocal microscopy, SUMO-2/3 was undetectable during completion of the first and second meiotic divisions (Fig. 3D, Meio) but was clearly localized to the chromocenters of forming round spermatids (RS, arrow), a pattern similar to that of SUMO-1 (Vigodner and Morris 2005).

Fig. 3
figure 3

Confocal microscopic localization of SUMO-2/3 in mouse spermatocytes. Nuclear SUMO-2/3 is not detectable in zygotene spermatocytes (Zy, panels A and A1, B and B1) but is concentrated in the sex body region in pachytene spermatocytes. (The insert in panel C is an isolated cell image where staining of centromeric heterochromatin is also evident.) At stages XII–I of mouse spermatogenesis (panel D), SUMO-2/3 is undetectable during meiotic divisions (Meio) but is concentrated in the chromocenters of the forming round spermatids (RS, arrows). Scale bar represents 10 µm

Full size image

Discussion

It had originally been suggested that the mechanism of MSCI was similar to X-chromosome inactivation and involved Xist RNA. However, later studies did not support this hypothesis, and revealed that MSCI occurs through different mechanisms (Turner 2007). MSCI has been studied intensively over the last few years. Although several major players in the process have been identified, questions still exist regarding the precise mechanism of MSCI. In order to better understand the role of new proteins in MSCI, the exact timing should be determined of their appearance over the XY chromosomes in relation to already identified molecular events. Although a widely used approach using surface-spread spermatocyte nuclei is helpful in differentiation of spermatocytes on the basis of the morphology of the synaptonemal complex, this method is limited in the detection of small changes within spermatocytes from the same development stage (e.g., leptotene) because (1) only a few of these cells can be found within the same microscopic field, and (2) unless a mutant animal model is used where a certain protein has been inactivated, it is difficult to determine which protein appears first and which second in the sequence of meiotic events, when these events progress rapidly. We have previously demonstrated that Confocal scanning laser microscopy is a valuable tool for studying spermatogenesis (Vigodner et al. 2002; Vigodner and Morris 2005). CLSM allows the study of spermatogenesis within intact seminiferous tubules under conditions that minimize cell and size deformation by avoiding the multistage process of paraffin section preparation. The major advantage over other methods for spermatocyte imaging using this technique is that numerous cells from the same developmental stage, such as leptotene or zygotene, can be visualized and analyzed without other germ cells present in the defined section. Furthermore, as the stages of spermatogenesis occur in sequence along the seminiferous tubules, sequential spermatocyte maturation and ordered appearance of different proteins may be followed by moving the laser beam along the length of the tubule.

In this study, we followed the localization of γH2AX and SUMO during the initiation of meiosis in intact seminiferous tubules, as studied by confocal microscopy. To shed more light on the mechanism of MSCI, we focused on early meiotic events. In leptotene spermatocytes, DNA double-strand breaks were marked by a γH2AX signal in the whole nucleus that intensified toward stages X–XI, a time point that may correspond to the maximal number of DNA double-strand breaks. It has been shown by Turner et al. (2005, 2007) that, during these stages, the X and Y chromosomes are still transcriptionally active. Their inactivation occurs only after the disappearance of γH2AX from the entire nucleus, and is coincident with its concentration over the sex chromosomes. The results of this study indicate that SUMO-1 appearance on the sex chromosomes precedes accumulation of γH2AX in the same nuclear, area and therefore occurs either before or very early during MSCI.

For reasons as yet undetermined, the proteins implicated in MSCI are also involved in the DNA damage response. Interestingly, during meiosis in Saccharomyces cerevisiae, sumoylation modifies Rad52 during the formation of DNA double-strand breaks, and also controls the assembly of the synaptonemal complex (Cheng et al. 2006; Sacher et al. 2006). This study did not detect SUMO at the sites of double-strand DNA breaks marked by the expression of γH2AX. Similarly, this and other studies were unable to detect SUMO along the synaptonemal complex in mouse spermatocytes, although a SUMO-modified enzyme was found associated with this meiotic structure. From these observations, it has been suggested that the SUMO level at certain stages may fall below the limits of microscopic detection (La Salle et al. 2008). Thus, a possible low presence of SUMO expression at the sites of DNA double-strand breaks cannot be excluded; this would be consistent with its role in MSCI, similar to those of other DNA repair proteins such as BRCA-1 and γH2AX.

We were unable to detect SUMO-2/3 in zygotene spermatocytes, but observed it in the sex bodies and paracentromeric heterochromatin of pachytene spermatocytes, sites where SUMO-2/3 has recently been shown by another group to localize, similarly to this study (La Salle et al. 2008). Such localization suggests that in contrast to SUMO-1, SUMO-2/3 might not be involved in MSCI but might have a role in the general events of heterochromatin formation in the sex body regions and at centromeres. The same function is also consistent with the presence of SUMO-2/3 in the chromocenters, regions of heterochromatic paracentromeric chromatin in round spermatids. These findings would be consistent with other studies that showed non-identical targets and both redundant and distinct functions of SUMO-1 and SUMO-2/3 within cells, although, as previously noted, a possible low and thus undetectable level of SUMO-2/3 in zygotene spermatocytes cannot be excluded. (Ayaydin and Dasso 2004; Vertegaal et al. 2004).

In conclusion, unlike SUMO-2/3, SUMO-1 is detectable on the sex chromosomes in zygotene spermatocytes, where it precedes γH2AX accumulation and MSCI. The specific targets of sumoylation and its regulation in spermatocytes are yet to be determined.