DGCR8 Localizes to the Nucleus as well as Cytoplasmic Structures in Mammalian Spermatogenic Cells and Epididymal Sperm

The localization of DGCR8 in spermatogenic cells and sperm from rat and mouse was studied by immunofluorescence and immunoelectron microscopy. Spermatogenic cells from these species yielded similar DGCR8 localization pattern. Immunofluorescence microscopy results showed thatDGCR8 localized to both the cytoplasm andnucleus. In the cytoplasm, diffuse cytosolic and discrete granular staining was observed. Dual staining showed that DGCR8 colocalized to the granules with MAEL (a nuage marker). In the nucleus of spermatocytes, both the nucleoli and nucleoplasm were stained, whereas in the nucleus of early spermatids small spots were stained. In late spermatids, DGCR8 localized to the tip of their head and to small granules (neck granules) of the neck cytoplasm. The neck granules were also observed in the neck of epididymal sperm. Immunoelectron microscopy results showed that DGCR8 localized to nuage structures. Moreover, DGCR8 localized to nonnuage structures in late spermatids. DGCR8 also localized to the nucleolus and euchromatin in spermatocytes and round spermatids and to small granules in the nucleus of late spermatids. The results suggest that in spermatogenic cells DGCR8 localizes not only to the nuclei but also to the cytoplasmic structures such as nuage and nonnuage structures. Furthermore, DGCR8 seems to be imported into the egg with neck granules in sperm during fertilization.


Introduction
Spermatogenesis is a dynamic and highly complicated process that is composed of three phases based on functional aspects: (1) the proliferative phase in which spermatogonia undergo rapid successive cell divisions, (2) the meiotic phase in which recombination and segregation of chromosomes occur, and (3) the differentiation phase, named spermiogenesis, in which spermatids transform into spermatozoa which are specialized to carry haploid genome to the egg [1][2][3]. During these phases, numerous phase-specific and housekeeping genes are expressed and supply proteins are required for phase-specific functions and usual metabolism in testis [4].
RNA silencing pathway plays important roles in spermatogenesis [5][6][7][8][9]. A key step in this pathway is the processing of double-stranded (ds) RNAs into short RNA duplexes of characteristic size and structures, which is performed by Drosha and Dicer, proteins related to the RNase III family [10][11][12][13]. DGCR8 is a protein encoded to the DiGeorge syndrome critical region gene 8 and a subunit of microprocessor complex. It has a nuclear transport signal sequence and usually located in the nucleus, especially in the nucleolus together with nucleolin [14]. In the nucleus, DGCR8 binds Drosha to form the microprocessor complex that cleaves the primary microRNA (pri-miRNA) to the hairpin-shaped pre-microRNA (pre-miRNA). The resulting pre-miRNA is then exported to the cytoplasm and further processed to mature miRNA by Dicer, which is loaded together with Ago2 proteins into the RNA-induced silencing complex (RISC). The miRNA guides RISC to silence target mRNAs through mRNA cleavage, translation repression, or deadenylation [10][11][12][13]. It is known that processing body (P-body) or stress granules are cytoplasmic sites for RISC functions [15][16][17][18][19].
In male germ cells, similar compartments are nuage structures, including chromatoid bodies, which contain several components involved in miRNA pathway [20][21][22]. This idea is expanded to "germ granules" [23], which widely exist in germ cell line [1,24]. We have attempted to clarify the nuclear localization of DGCR8 in spermatogenic cells to know the localization of DGCR8 in differentiating spermatogenic cells. The results showed that DGCR8 localized not only to the nucleus but also to diverse cytoplasmic structures. In this study, therefore, we investigate the precise localization of DGCR8 in spermatogenic cells and epididymal sperm of mammals using immunofluorescence (IF) and immunoelectron microscopy (IEM) techniques.

Animals.
Male Wistar rats (weight 180-220 g) and mice (25 g) were purchased from Kyudo Co. Ltd. (Tosu, Japan). All animals received standard diets and water ad libitum until use. Health epididymides of pig and horse were obtained from a slaughterhouse. Sperms were collected from epididymides of rat, mouse, pig, and horse. An experiment was performed in accordance with the guidance for Animal Experiments issued by the Nagasaki International University.

Western Blotting.
Testes of rat and mouse (250 mg each) were homogenized in 5 mL of sample buffer for SDSPAGE. Homogenates were heated in boiling water for 5 min and centrifuged at 10,000 ×g for 30 min. The resulting supernatants were divided into 100 mL aliquots and stored at −80 ∘ C. Protein concentration of 5% homogenates (w/v) in PBS was determined by the bicinchoninic acid method (Pierce Chemical, Rockford, IL, USA) using bovine serum albumin as a standard. Fifty micrograms of each sample were analyzed by Western blotting. The apparent molecular mass of DGCR8 was estimated by a prestained protein maker (Nippon Genetics Europe, Düren, Germany).

Immunofluorescence Staining.
Smear preparations of epididymal sperm from rat, mouse, pig, and horse, and frozen sections (6 m thick) of testes from rats and mice, were fixed in 4% paraformaldehyde (w/v) in 0.1 M Hepes-KOH buffer (pH 7.4) for 15 min. After permeation treatment with 0.1% Triton X-100 (v/v) + 0.2% Saponin (w/v), smear preparations and sections were incubated in 2% fish gelatin (w/v) in PBS for 30 min to block nonspecific adsorption of IgG, followed by overnight incubation with anti-DGCR8 antibodies from ProteinTech (×100) or from Abnova (×500 by). After washing with PBS, the reacted IgG was visualized by Alexa 568 or Alexa 488-conjugated goat anti-rabbit IgG. For dual staining, sections were incubated with both rabbit anti-DGCR8 antibodies and affinity-purified guinea pig anti-MAEL (1 g/mL) and each reacted IgG was visualized by Alexa 568-conjugated goat anti-rabbit IgG and Alexa 488-conjugated goat antiguinea pig IgG. For control, nonimmune sera were used instead of the specific primary antibodies. Nuclei were stained by 3 M DAPI (Hoechst, Tokyo, Japan) for 60 min at RT. The smear preparations and sections were examined with a Nikon Eclipse E600 fluorescence microscope (Tokyo, Japan). The images were merged with Adobe Photoshop 7.0 to determine whether each antigen colocalizes to the same area. The stage of the seminiferous cycle was determined from the localization of elongating and elongated spermatids with individual tubules as previously described [27].

Immunoelectron Microscopy.
Rat and mouse testes were cut into small tissue blocks in fixative consisting of 4% paraformaldehyde (w/v), 0.2% glutaraldehyde (v/v), 0.02% CaCl 2 (w/v), and 0.1 M Hepes-KOH (pH 7.4) and tissue blocks were kept in the fixative for 1 h at 4 ∘ C. Tissue blocks were dehydrated in ethanol and embedded in LR White at −20 ∘ C, followed by resin polymerization under UV light at −20 ∘ C. Thin sections were cut with a diamond knife equipped with a Reichert Ultracut R, mounted onto nickel grids, and incubated with rabbit anti-DGCR8 antibodies from ProteinTech (×100) or Abnova (×500) overnight at 4 ∘ C. For control experiments, nonimmune rabbit serum was used instead of the specific primary antibodies. The reacted IgG was visualized by 15 nm protein A-gold probe. Sections were contrasted with uranyl acetate and lead citrate, coated with evaporated carbon and examined with a Hitachi H7650 electron microscope (Tokyo, Japan) at an acceleration voltage of 80 kV. The stage of the seminiferous cycle and step of spermatids were determined as described by Russell et al. [27].

Routine Electron Microscopy.
Testis tissue blocks from rats and mice were fixed in the fixative containing 4% paraformaldehyde (w/v), 1% glutaraldehyde (v/v), 0.02% CaCl 2 (w/v), and 0.05 M Hepes-KOH (pH 7.4) overnight at 4 ∘ C. After brief wash in PBS, testis tissue blocks were fixed in 1% reduced osmium tetroxide (w/v) for 1 h, dehydrated, and embedded in Epon. Thin sections were contrasted with lead citrate and examined with electron microscope.

Western Blotting.
Two protein bands of 117 and 86 kDa were observed in testis homogenates of rat and mouse with rabbit anti-DGCR8 antibody from ProteinTech ( Figure 1). The same results were obtained with antibody from Abnova (data not shown). The band of 86 kDa was consistent with the molecular weight of rat DGCR8 shown in SWISS-PROT data base.

Immunofluorescence Staining.
Immunofluorescence signals for DGCR8 were observed in three different compartments such as cytoplasmic granules, cytosol, and nuclei. Staining patterns changed during the stages cycle of seminiferous tubules. We observed carefully the staining pattern during spermatogenesis of rat testis. Cytoplasmic granular staining in spermatids was observed at steps 1-17 ( ). In control, the staining described above was completely eliminated (Figure 2(f)).
Next we stained dually testis sections for DGCR8 and MAEL, a nuage protein marker, in order to clarify whether DGCR8-positive cytoplasmic granules are nuage. The results are shown in Figure 3. DGCR8-positive granules in round spermatids were stained also for MAEL, whereas granules in the cytoplasm of late spermatids were negative for MAEL (Figures 3(a)-3(c)). In late spermatocytes, MAEL-positive granules were not stained for DGCR8. In the cytoplasm of step 11-15 spermatids, granules stained for both DGCR8 and MAEL were observed (Figures 3(d)-3(f)). In some spermatocytes at stages V-VI, large spot within the nucleus was heavily stained for DGCR8 but not for MAEL ( Figures  3(d)-3(f)).

Immunoelectron Microscopy.
Similar results were obtained with anti-DGCR8 antibodies from ProteinTech and Abnova.

Nuage Structures.
We studied whether DGCR8 localized to nuage structures of rat and mouse, which previously we classified into five types [28]. Gold labeling was generally weaker than that for the other nuage proteins studied previously [25,[29][30][31] irregularly shaped perinuclear granules, intermitochondrial cement, and chromatoid bodies were labeled for DGCR8 (Figures 4(a)-4(f)). 70-90 nm particles and satellite bodies were not stained for DGCR8 (data not shown). No gold labeling in these structures was observed in control (Figure 4 insets).

Nonnuage Structures.
Mitochondria-associated granules, which appeared in the elongated cytoplasmic lobe of late spermatids, were stained for DGCR8 (Figures 5(a) and 5(b)). Ribosome aggregates, which emerged in the residual body of step 19 spermatids, were moderately labeled ( Figures  5(c) and 5(d)). In addition, gold labeling for DGCR8 was observed in the small granules which were aggregated in the cytoplasm of step 17 spermatids and in the residual body ( Figures 5(e) and 5(f)). The other nonnuage structures such as granulated bodies and reticulated bodies were negative for DGCR8 (data not shown). No gold labeling was seen in IEM control sections (Figure 5 Insets).

Nucleoplasm.
The nucleoplasm of pachytene spermatocytes was moderately labeled but that of spermatids very weak or not. In spermatocytes, especially in late spermatocytes, most of the labeling was associated with euchromatin ( Figure 6(a)). In step 6-7 spermatids, small dense particles, which aggregated to form a cluster, were labeled for DGCR8, whereas other nucleoplasm area was not labeled (Figure 6(b)). The nucleoplasm of step 10-11 spermatids was not stained excepting aggregated small dense granules (Figure 6(c)). No gold labeling was observed in IEM control sections (data not shown).

Nucleolus.
Gold labeling for DGCR8 in the nucleolus was weaker in early to middle pachytene spermatocytes and became remarkable after stage IX, in which the nucleolus frequently came into close contact with XY body. In the nucleolus, DGCR8 mainly localized to the granular component surrounding the fibrillar center to which no DGCR8 localized (Figure 7(a)). Dense fibrillar component was also negative for DGCR8 (Figure 7(a)). In the nucleolus closely associated with XY body, DGCR8 was detected in both the granular component and in the fibrillar center (Figure 7(b)), whereas the dense fibrillar component was not labeled. The XY body was labeled very weak or not (Figure 7(b)). No gold labeling was observed in IEM control sections (data not shown).

Dense Material in the Cytoplasm of Sertoli Cell Ectoplasmic Specialization.
Ectoplasmic specializations are the complex cytoskeletal structure composed of the Sertoli cell plasma membrane closely contacted with spermatid head, the endoplasmic reticulum (ER) of this cell, and actin bundles between the plasma membrane and the ER [32]. DGCR8 localized to dense material located between the plasma membrane and the ER (Figure 8(a)).

Small Granule in the Neck Cytoplasm of Late Spermatids.
This granule consisted of dense material, appeared in spermatids after step 9, and could be observed also in the epididymal sperm. It existed solitarily or attached to the neck piece of the flagellum. We named it neck granules. DGCR8 localized to the neck granules (Figures 8(b) and 8(c)).
Although the electron density of cross-sectioned annulus was very similar to that of the neck granules, no DCGR8 was detected in it (Figures 8(b) and 8(c)). No gold signals were seen in IEM control section (Figure 8(d)). The neck granules were evidently confirmed in routine electron microscopy ( Figure 8(e)).

Neck Granules in Epididymal
Spermatozoa. Next, we confirmed whether the neck granule existed in the neck cytoplasm of epididymal sperm from several mammals. Neck granules in sperm from rat, mouse, pig, and horse were stained for DGCR8 (Figures 9(a), 9(c), 9(d), and 9(e)). In rat and mouse, the number of stained granule was one or two, whereas in pig and horse, more than two granules were frequently stained (Figures 9(d) and 9(e)). No granules were stained in IF control preparation (Figure 9(b)). weight estimated from amino acid sequence of DGCR8 cited in SWISS-PROT database. The molecular weight of the slowly migrated band was estimated to be 117 kDa. It has been reported that the nuclear localization sequence (NLS) of DGCR8 is 1-275 amino acid sequence [33] and later confined to be 256-289 sequence [14]. According to manufacturer datasheet, the protein with 117 kDa (120 kDa by manufacturer) is an isoform of DGCR8 but not well characterized so far. In the present study, these protein bands were also observed in Western blotting with anti-DGCR8 antibody from Abnova. Therefore, it is likely that 117 k Da protein is a modified larger isoform of 86 kDa protein, which has DGCR8 epitope(s). If so, both 86 and 117 kDa proteins have the NLS and are able to localize to the nucleus. However, it is not clear which proteins localize to the cytoplasmic structures such as nuage and other structures investigated in this study.

Nuclear Localization of DGCR8 in Spermatogenic Cells.
We showed that in spermatogenic cells DGCR8 localized not only to the nuclei but also to the cytoplasmic compartments, including nuage and nonnuage structures. Our IF results on the nuclear localization of DGCR8 are similar to those reported previously [14,33]. The staining intensity of DGCR8 Step 7 Step 10 (c) in spermatogenic cells changed during differentiation of spermatogenic cells. In pachytene spermatocytes, the strongest staining was observed at around stage XIII. This suggests that the cells are preparing miRNA silencing pathway against transposon appearing in meiosis shortly after this stage [34][35][36]. Our IEM study clarified that in the nucleolus DGCR8 was almost exclusively localized to granular component but not dense fibrillar component. The results are similar to previous report [14]. The nucleolus disappears in late spermatids. In such cells, DGCR8 was found to be associated with the aggregate of small dense particles in the nucleoplasm. The nature of these particles is unclear. Electron density of them is very similar to that of ribonucleoproteins. to various cytoplasmic structures, including nuage and nonnuage structures specific for spermatogenic cells. The nuage is known to be a cytoplasmic site for RNA processing, including translation initiation [37,38], RNA silencing [35,39], RNA decay [40,41], and RNA storage [23,42]. The function of DGCR8 in nuage is not known. The localization of DGCR8 in the nuage suggests that it has some sort of function which is not relevant to serving as a subunit of microprocessor complex. In the present study, we showed that DGCR8 was present in nonnuage structures such as mitochondriaassociated granules, ribosome aggregates, and small granules in the residual bodies. We have revealed that several nuage proteins are also contained in these structures [25,28,29,31], to which DGCR8 was added as a constituent protein in this time. However, the role of the nonnuage structures in spermatogenesis is quite unknown. This is our future problem to solve. Interesting finding is the localization of DGCR8 in neck granules of late spermatids and epididymal sperm. In addition, the granules were confirmed to exist in sperm of pig and horse. The results suggest that DGCR8 is a constituent protein in the neck granules of mammals. Previously, we have shown that MAEL protein, which plays a central role as transposon silencer during spermatogenesis, localizes to neck granules [25]. The relationship between these proteins is not obvious. It is interesting that DGCR8 might be carried into egg by the neck granules during fertilization.

Conclusion
DGCR8 localizes not only to the nucleus but also to the cytoplasmic structures, including nuage and nonnuage structures in spermatogenic cells. In the nucleus, DGCR8 is mainly present in nucleolus, especially in the granular component. In the nuage, DGCR8 localizes to irregularly shaped perinuclear granules, intermitochondrial cement, and chromatoid bodies. DGCR8 localizes to the neck granules of late spermatids and epididymal sperm.