Requirement for Sun1 in the expression of meiotic reproductive genes and piRNA

Abstract

The inner nuclear envelope (NE) proteins interact with the nuclear lamina and participate in the architectural compartmentalization of chromosomes. The association of NE proteins with DNA contributes to the spatial rearrangement of chromosomes and their gene expression. Sun1 is an inner nuclear membrane (INM) protein that locates to telomeres and anchors chromosome movement in the prophase of meiosis. Here, we have created Sun1-/- mice and have found that these mice are born and grow normally but are reproductively infertile. Detailed molecular analyses showed that Sun1-/- P14 testes are repressed for the expression of reproductive genes and have no detectable piRNA. These findings raise a heretofore unrecognized role of Sun1 in the selective gene expression of coding and non-coding RNAs needed for gametogenesis.

Fig. 1.

Defective gametogenesis in Sun1^–/– mice. (A) Schematic representation of the wild-type allele, the targeting vector and the mutated locus. The targeting vector contains the PGK-Neo gene (Neo) and the thymidine kinase gene (TK). Three loxP sites (denoted by black triangles) were placed at the 5′ end of Neo, between Neo and exon 10 (E10) and at the 3′ end of exon 11. Sun1 was removed by crossing mice carrying the Sun1 targeting vector with whole-body Cre transgenic mice. Gray arrows indicate the relative positions of primers used for genotyping by PCR. (B) PCR analysis of representative Sun1 offspring from heterozygous matings. PCR of wild-type genomic DNA generates a 1262 bp fragment, whereas the targeting vector (left) generates a 2570 bp fragment. The sequence between the loxP sites was removed after Cre induction (right), and a 263 bp fragment was generated. (C) Comparison of testes from 4-week- and 7-month-old Sun1^+/+ and homologous mutant Sun1^–/– mice. Testes from Sun1^–/– mice are smaller in both cases. Clear weight differences were observed at about 4-weeks post birth. (D) Hematoxylin and Eosin (H&E)-stained sections (200× and 1000× magnifications) of testis from 4-week-old mice. In Sun1^+/+ testis, a clear progression of the first wave of spermatogenesis with differentiated spermatids (sd) was observed. No spermatid was found in the Sun1^–/– mice. Instead, accumulation of zygotene-like spermatocytes was frequently detected in the Sun1^–/– seminiferous tubules. sg, spermatogonia; st, sertoli cells; sc, spermatocyte; sd, spermatid.

Fig. 2.

Comparison of γH2AX staining in Sun1^+/+ versus Sun1^–/– mice. Distribution of γH2AX (red, Alexa-594) in Sun1^+/+ (A-O) and Sun1^–/– (P-U) testes at meiosis prophase I using spermatocyte spread (A-L,S-U) or cryosection (M-R). SCP3 (green, Alexa-488) staining denotes the stage of meiosis. A-L show single wild-type cells stained with γH2AX and SCP3. Overviews of testis cells (cryosections) in Sun1^+/+ (M-O) and Sun1^–/– (P-R) are shown. A massive accumulation of γH2AX-stained cells was observed in Sun1^–/– (P-R) but not in Sun1^+/+ (M-O) mice. In Sun1^+/+ leptotene spermatocytes, γH2AX was distributed all over the chromosome (panels 1-3). From zygotene to pachytene, γH2AX staining decreased gradually, and finally was restricted to the gonosomal chromatin in late zygotene/early pachytene spermatocytes (G-L). By contrast, Sun1^–/– cells (spermatocyte spread) showed a general γH2AX staining pattern, similar to zygotene spermatocytes (P-U). Arrows in K and N denote the XY body.

Fig. 3.

Localization of Sun1 in the wild-type testis. (A) Immunofluorescent staining of Sun1 in mouse embryonic fibroblasts (MEFs) using αmusSun1-C. A clear nuclear envelope localization of Sun1 (green, Alexa-488) was seen. DNA stained by DAPI is in red. (B) Localization of Sun1 in the wild-type adult testis (Davidson's fixed, paraffin-embedded tissue section), as detected by immunohistochemistry. Sun1 is stained brown, and the nuclei were counterstained with Methyl Green (green). Part a shows an overall view of Sun1 staining in seminiferous tubule. Sun1 was localized to the nuclear periphery in spermatogonia (b) and round spermatids (g), showed punctate staining at chromosome ends in prophase I (c-f), and decorated an acrosome-like structure in elongated spermatids (h). Sun1 staining was not seen in spermatozoa (i). (C) Immunofluorescent staining of a frozen testis section (28-day-old) by αmusSun1-C (green, Alexa-488), anti-Trf1 (red, Alexa-594) and anti-lamin B1 (gray, Alexa-633) antibodies. Nuclei are labeled by DAPI (blue). Insets show higher-magnification images of the framed sections. sc, spermatocyte; sp, spermatozoan; Esd, elongated spermatid.

Fig. 4.

Histology of testes from Sun1^+/+ and Sun1^–/– mice during the first wave of spermatogenesis. (A) A schematic time line for the first wave of mouse spermatogenesis. (B) H&E stainings of testes from 9- and 14-day-old Sun1^+/+ and Sun1^–/– mice are shown. Insets show higher-magnification images of the tissues. In P9 testes (a,b), no significant morphological difference was observed between Sun1^+/+ and Sun1^–/– mice. In P14 mice, although the distribution of spermatogonia/sertoli cells and meiotic prophase I spermatocytes are similar (quantified in C) in the wild-type and Sun1^–/– testes, wild type mostly showed clear `bouquet' structures (c) while Sun1^–/– cells did not (d). Cell stages later than pachytene were not observed in either P9 or P14 testes. (C) Cell compositions in P9 and P14 Sun1^+/+ and Sun1^–/– testes. Two thousand cells were counted in each set.

Fig. 5.

Reduced Mili- and Miwi-associated piRNA expression in Sun^–/– germ cells. (A) Semi-quantitative RT-PCR was used to validate the expression of spermatogenesis-related genes at P9, P14 and P28 testes. Gapdh was used for normalization. (B) Radioactive labeling of total RNAs isolated from 28-day-old mouse testis. Sun1^+/+ mice showed an abundant ∼30-nucleotide small-RNA fraction while no detectable ∼30-nucleotide small RNAs were seen in Sun1^–/– mice. M, radio-labeled RNA ladder. (C) Northern blot hybridization for Mili- or Miwi-associated piRNAs. Ten micrograms of total mouse RNA were loaded in each lane. A mixture of probes against Mili-associated piRNAs from chromosome 9 and 17 or Miwi-associated piRNAs (piR-1, piR-2, piR-3) were used. A probe for microRNA miR-16 was hybridized simultaneously as a control. Blots were re-probed with a U6 antisense probe for comparison. (D) Semi-quantitative analysis of individual Mili- and Miwi-associated piRNA by RT-PCR. Expression of miR-16 is shown for normalization. C, control lane (no RNA was loaded in the reaction). (E) Semi-quantitative RT-PCR analysis of Line-1 type A element in Sun1^–/– and Sun1^+/+ P14 testes. Gapdh expression was used as a control. Equal amounts of total RNA were used for RT-PCR cycles (25, 30 and 35 cycles from left to right, respectively).