MicroRNAs downregulate Bag of marbles to ensure proper terminal differentiation in the Drosophila male germline

Adult stem cells sustain tissue homeostasis by consistently providing differentiated cells lost by injury or turnover (Morrison and Spradling, 2008). To retain an adequate stem cell population while generating differentiated cells, many adult stem cells asymmetrically divide into two daughter cells. One daughter cell self-renews as a new stem cell and the other daughter cell differentiates (Knoblich, 2008). In order to generate a sufficient number of differentiated cells in many adult stem cell lineages, the daughter cells from a stem cell asymmetric division undergo a proliferative stage called transit-amplification before activating a gene expression program required for terminal differentiation (Clarke and Fuller, 2006).

The Drosophila melanogaster germline stem cell (GSC) lineages have been used to study molecular mechanisms that determine the transition from the transit-amplification stage to the differentiation stage. In Drosophila testes, a GSC asymmetrically divides to produce a self-renewed GSC and a gonialblast, which undergoes exactly four rounds of mitosis as transit-amplifying spermatogonial cells. The resulting 16 spermatogonial cells then enter the spermatocyte stage, in which a robust transcription program is activated to express genes for meiotic divisions and terminal differentiation (Fuller, 1993; White-Cooper, 2010). Throughout spermatogenesis, germ cells divide and differentiate in a syncytium. It is only during the late stage of spermiogenesis that the cytoplasm of spermatids is separated by a process called individualization, in order to produce individual and functional haploid sperm (Fuller, 1993; Lindsley and Tokuyasa, 1980).

The transition from the transit-amplifying stage to the terminal differentiation stage should be tightly controlled. Failure of such control may lead to tumor-containing, ever-dividing transit-amplifying cells or to a lack of terminally differentiated cells, which leads to tissue dystrophy or to infertility. In the Drosophila male germline lineage, bag of marbles (bam) is essential for the proper transition from the proliferative to differentiating spermatocyte stages (Gönczy et al., 1997; McKearin and Spradling, 1990). Bam is detected in four- to 16-cell spermatogonia, with levels peaking at eight-cell spermatogonia (Gönczy et al., 1997). It has been shown that accumulation of Bam to a threshold level is required for the transition from spermatogonia to spermatocytes (Insco et al., 2009). However, Bam is not detected in GSCs (Insco et al., 2009; Monk et al., 2010; Schulz et al., 2004). If Bam is ectopically expressed in GSCs, it causes GSC premature differentiation or cell death, leading to GSC loss (Schulz et al., 2004; Sheng et al., 2009). Therefore, expression of Bam needs to be tightly controlled during spermatogenesis.

We report that in the Drosophila male GSC lineage, bam mRNA is present but Bam is undetectable in spermatocytes. Repression of Bam protein accumulation is achieved by microRNAs via the bam 3′UTR. Failure to repress Bam protein accumulation in spermatocytes leads to spermiogenesis defects and male sterility. Our results demonstrate endogenous functions of microRNAs in controlling stem cell differentiation pathway through regulating a key differentiation factor. Although Bam is important for initiating GSC differentiation, our data show that its downregulation is crucial for proper spermatid terminal differentiation.

Flies were raised on standard molasses medium at 25°C. Wild-type flies were y,w. The sa-GFP (Chen et al., 2005), nos-Gal4 (Kiger et al., 2001) and bam-GAL4 (Chen and McKearin, 2003b) flies have been described previously. The bam^Δ86 (BL5427), Df(3R)FDD-0247627 (Deficiency line for bam, BL27401), UAS-GFP (BL4775), UAS-Gal4 (BL5939) and dcr-1^Q1147X (BL32066) flies were obtained from Bloomington Stock Center. The hs-FLP; Act5c-FRT-stop-FRT-Gal4, UAS-GFP (Secombe et al., 2007) were obtained from J. Secombe (Albert Einstein College of Medicine, Bronx, NY, USA). The following stocks were generated in this study: bam-HA bam3′, bam-HA tub3′, bam-HA with the second half of bam 3′UTR substituted with α-tubulin84E 3′UTR, bam-HA with the first half of bam 3′UTR substituted with α-tubulin84E 3′UTR, bam-HA bam3′ with the putative miR-275/miR-306-binding site mutated, bam-HA bam3′ with the putative miR-7-binding site mutated and bam-HA bam3′ with both putative miR-317-binding sites mutated (supplementary material Table S1). All seven constructs described above were incorporated at the attP40 site as transgenes using the ϕC31 integrase (Groth et al., 2004). UAS-miR-7, UAS-miR-275, UAS-miR-306, UAS-miR-317 (supplementary material Table S2), UAS-GFP-miR-275SP and UAS-GFP-miR-306SP (supplementary material Tables S3 and S4) flies were generated in this study and the transgenes were introduced into flies by P-element-mediated transformation via embryo injection.

For efficient SP expression, bam-Gal4, UAS-Gal4/UAS-GFP-microRNA-SP; dcr-1^Q1147X /UAS-GFP-microRNA-SP flies were raised at 29°C. For the fertility test, newly eclosed single males were mated with 2 y,w virgin females. Sterility was determined if there was no crawling larvae 5-7 days after the crosses. Testes from sterile flies were dissected and the testis morphology was examined using phase-contrast microscopy.

In situ hybridization was performed as previously described (Morris et al., 2009). DIG-labeled antisense bam probes were synthesized from PCR products amplified using a set of primers (forward 5′-ACCAGCAGTTGGACCACAAT-3′ and reverse 5′-GGATCCTAATACGACTCACTATAGGGAGAATGCGCAGCCTTGATCCAAT-3′). The probes miR-275 (33020-01) and miR-306 (33047-01) were purchased from EXIQON. Images were taken with a Zeiss Axioplan microscope.

Clones co-expressing UAS-miR275 or UAS-miR306 with the GFP marker were generated using flies with the following genotypes: hs-FLP; Act5c-FRT-stop-FRT-Gal4, UAS-GFP/UAS-miR275, UAS-miR275/+ or hs-FLP; Act5c-FRT-stop-FRT-Gal4, UAS-GFP/UAS-miR306; UAS-miR306/+. Clones were induced by heatshock treatment of pupae at day 8 and day 9 for 2 hours at 37°C each time. After the second heat shock, flies were placed in 25°C incubator for 3 days, followed by dissection for immunostaining.

Immunostaining was performed as previously described (Insco et al., 2009). Primary antibodies were rat anti-HA (Roche, 3F10) at 1:50, rabbit anti-Vasa (Santa Cruz, sc-30210) at 1:200, chicken anti-GFP (Abcam, ab13970) at 1:1000, mouse anti-Fas3 (DSHB, 7G10) at 1:100, mouse anti-Bam (DSHB) at 1:20 and mouse anti-Lamin C (DSHB, LC28.26) at 1:20. Secondary antibodies were the Alexa Fluor-conjugated series (1:200; Molecular Probes). For phalloidin staining, testes were dissected in PBS and fixed in 1 ml 4% formaldehyde in PBS for 30 minutes at room temperature. After fixation, testes were rinsed twice with 0.1% Triton X-100 in PBS (PBST) for 10 minutes at room temperature and then incubated with Alexa Fluor 568 phalloidin (Invitrogen, A12380) in the dark for 2 hours at room temperature. Testes were rinsed again twice with PBST, for 10 minutes each time, and mounted with Vectashield medium (Vector H-1200). Images were taken using the Zeiss LSM 510 Confocor 3 NLO system and processed using Adobe Photoshop.

Oligonucleotide primers were designed for bam (forward 5′-ACTCAGCGCATGGAGAGATTGCTA-3′ and reverse 5′-AGTAGCGGTGCTCCAGATCCATTT-3′) and a housekeeping gene, RpL32 (forward 5′-CATGCTGCCCACCGGATTCAAGAAG-3′ and reverse 5′-CTCGTTCTCTTGAGAACGCAGGCGA-3′). Testes cDNA preparation and RT-PCR were performed as described previously (Chen et al., 2011). Each PCR reaction was performed in duplicates and Ct numbers were averaged, at least three independent biological replicates were used. The relative expression level was calculated by 2^{Ct(Rpl32)–Ct(bam)}, where Rpl32 was used as an internal control.

Testes were dissected in PBS and cut at about half way from the testis tip using forceps. Then, the testes tips were transferred onto a microscope slide and covered with a coverslip. PBS on the slide was carefully drained using a Kimwipe until cysts were flattened to single-cell-layer. The number of cells in spermatocyte cyst was counted using phase contrast microscopy. Images were taken on a Zeiss Axioplan microscope using the Openlab software.

Potential microRNA binding sites at bam 3′UTR were identified using mirBase (http://www.mirbase.org/).

Bam protein is detected in 4-16 spermatogonial cells with a peak expression at the eight-cell spermatogonial stage (Fig. 1A) (Gönczy et al., 1997; Insco et al., 2009). To identify the bam mRNA pattern, in situ hybridization was performed using antisense bam probe in wild-type testes. Two waves of bam transcript were detected. The first wave was in the spermatogonial cells (Fig. 1B, single asterisk) and the second wave was in early spermatocytes (Fig. 1B, double asterisk). As controls, such specific signals were not detected using sense bam probe in wild-type testes (supplementary material Fig. S1A) or using antisense bam probe in transcript null bam^Δ86/Df testes (Gönczy et al., 1997; McKearin and Ohlstein, 1995) (supplementary material Fig. S1B). The presence of bam mRNA, but absence of Bam protein, indicated that accumulation of Bam protein in spermatocytes is repressed post-transcriptionally.

The 3′UTR is a common recognition site of post-transcriptional regulators (Lai, 2002; Stark et al., 2005). To investigate whether bam is regulated post-transcriptionally via its 3′UTR, the bam-HA genomic construct (including the endogenous bam promoter, 5′UTR, coding sequence tagged with HA at the 3′ end and 3′UTR), which rescues the bam-null phenotype in both male and female gonads (Li et al., 2009), was used to generate transgenic flies. In addition, the bam 3′UTR sequence was replaced with the 3′UTR of a constitutively expressed α-tubulin84E gene (Gan et al., 2010). Then, both bam-HA with its endogenous 3′UTR (bam-HA bam3′) and the one with a-tubulin84E 3′UTR (bam-HA tub3′) transgenes were integrated into fly genome using the φC31 integrase-mediated site-directed method in order to ensure the expression of both transgenes in the same genomic context (Groth et al., 2004). In order to distinguish spermatogonia and spermatocytes precisely, we used a spermatocyte-specific cellular marker, Sa-GFP (Chen et al., 2005) (Fig. 1C,D). Immunostaining experiments showed the presence of Bam in spermatogonial cells in testes from bam-HA bam3′ males (Fig. 1C), which is consistent with previous reports (Gönczy et al., 1997; Insco et al., 2009). By contrast, Bam was present in both spermatogonia and spermatocytes in testes from bam-HA tub3′ males (Fig. 1D). These results demonstrate that the 3′UTR of bam is required for the repression of Bam protein accumulation in spermatocytes.

Conversely, Bam needs to be silenced in both male (Fig. 2A) and female (Fig. 2C,C′) GSCs to prevent precocious differentiation (Ohlstein and McKearin, 1997; Schulz et al., 2004; Sheng et al., 2009). However, substitution of the bam 3′UTR with the α-tubulin84E 3′UTR did not lead to ectopic Bam expression in either male (Fig. 2B) or female (Fig. 2D,D′) GSCs, suggesting that the absence of Bam in GSCs is most likely to be regulated transcriptionally, consistent with previous reports (Chen and McKearin, 2003a; Chen and McKearin, 2003b).

In Drosophila ovaries, Bam is detected in cystoblasts, the immediate daughter cells of female GSCs, which will undergo differentiation, as well as in young developing cysts. However, it is not detected from germarium region 2a and beyond (Fig. 2C,C′) (Ohlstein and McKearin, 1997). In contrast to the results in testes (Fig. 1D), persistent bam expression was not detected in later stage germ cells in bam-HA tub3′ ovaries (Fig. 2D,D′). Together, these data reveal that the post-transcriptional regulation of Bam via the bam 3′UTR to prevent accumulation of Bam in later-stage germ cells is male specific.

Although post-transcriptional regulation can be achieved by mRNA degradation or by translational repression, these two mechanisms are not mutually exclusive (He and Hannon, 2004; Hesketh, 2004). To study whether the endogenous bam 3′UTR is required for downregulating bam mRNA, quantitative RT-PCR was performed to compare the bam transcript level in testes from bam-HA bam3′ and bam-HA tub3′ transgenic males, including both transgenic and endogenous bam transcript. We found that levels of bam mRNA were 2.4-fold in bam-HA bam3′ testes and 3.7-fold in bam-HA tub3′ testes, respectively, compared with wild-type testes (Fig. 3A). The higher level of bam transcript in bam-HA tub3′ testes than in bam-HA bam3′ testes suggest that the bam 3′UTR could be used to destabilize bam mRNA.

When Bam is stabilized in spermatogonia, previous studies have demonstrated that the transition from spermatogonia to spermatocyte occurs prematurely, leading to spermatocyte cysts with fewer than 16 cells (Insco et al., 2009). Therefore, we next studied whether the bam 3′UTR-dependent post-transcriptional regulation of Bam affects the spermatogonia-to-spermatocyte transition. We counted the number of spermatocytes in intact cysts from either bam-HA bam3′ or bam-HA tub3′ testes, using both phase-contrast microscopy (Fig. 3B,C) and the spermatocyte-specific Sa-GFP marker (Fig. 3B′,C′). We found that the percentage of testes containing at least one eight-cell spermatocyte cyst had no significant difference between these two strains (Fig. 3D, P=0.93), indicating that both transgenes behaved similarly in regulating the spermatogonia-to-spermatocyte transition.

MicroRNAs commonly act through 3′UTRs of target transcripts to destabilize mRNAs or repress translational initiation (He and Hannon, 2004). Bioinformatics analysis was performed to screen for putative microRNA recognition sites at the bam 3′UTR sequences (Fig. 4A). Four putative microRNA-binding sites were identified: miR-7, miR-275, miR-306 and miR-317, including one site shared by miR-275 and miR-306, and two sites recognized by miR-317 (Fig. 4A). Furthermore, all four sites were located at the second half of the bam 3′UTR (sequences not highlighted in Fig. 4A). By contrast, no microRNA-binding site was found at the α-tubulin84E 3′UTR.

To investigate whether the second half of the bam 3′UTR containing all putative microRNA recognition sites is required for downregulation of Bam protein in spermatocytes, α-tubulin84E 3′UTR was used to specifically replace the second half of bam 3′UTR so that the entire 3′UTR length would be similar to the endogenous bam 3′UTR (Fig. 4B). With this substitution, Bam expression was detectable in spermatocytes (Fig. 4B), indicating that the second half of bam 3′UTR was necessary for downregulation of Bam in spermatocytes. By contrast, when only the first half of the bam 3′UTR was replaced by α-tubulin84E 3′UTR, the bam expression pattern was unchanged (Fig. 4C), suggesting that the second half of bam 3′UTR was sufficient to downregulate Bam in spermatocytes. In summary, these results suggest that the downregulation of Bam in spermatocytes most probably occurs through microRNAs.

In animals, microRNAs usually bind to their target mRNAs through seed sequences, which are 2-7 nucleotide mRNA sequences complementary to the 5′ sequences of microRNAs (He and Hannon, 2004; Hesketh, 2004). Mutation of seed sequences can abolish the binding between microRNAs and their target mRNAs (Bartel, 2009). To pinpoint which putative microRNA-binding site is necessary for post-transcriptional regulation of Bam, seed sequences of each putative microRNA-binding site (Fig. 4A) were replaced with a 6 bp restriction enzyme site. We found that mutation of the seed sequence shared by miR-275 and miR-306 led to abundant Bam accumulation in spermatocytes (Fig. 5A). By contrast, mutations of seed sequences of either the miR-7- or the miR-317-binding site did not change Bam endogenous pattern (Fig. 5B,C). These results suggest that miR-275 and/or miR-306 are the potential trans-regulatory factors that downregulate Bam in spermatocytes.

Next, in order to examine whether miR-275 and/or miR-306 represses Bam accumulation, we used a sensitive assay for Bam accumulation. Each of the microRNAs was overexpressed using an early germ cell-specific nos-Gal4 driver (Van Doren et al., 1998). Thus, if overexpression of microRNAs is sufficient to block Bam protein accumulation, a delayed transition from spermatogonia to spermatocytes would be observed using a single spermatocyte cyst-counting assay. Indeed, we found that overexpression of either miR-275 or miR-306 resulted in testes containing 32-cell spermatocyte cysts (Fig. 5D,E), suggesting an extra spermatogonial division before switching to spermatocytes (Insco et al., 2009). By contrast, neither miR-7 nor miR-317, when overexpressed using the same nos-Gal4 driver, led to testes containing 32-cell spermatocyte cysts (Fig. 5E). Moreover, addition of one copy of bam-HA transgene fully rescued the 32-cell spermatocyte cysts phenotype in miR-275- or miR-306-overexpressing testes (Fig. 5E), confirming that this phenotype was caused by reduced Bam levels. In addition, we found that Bam was downregulated in miR-275- or miR-306-overexpressing spermatogonial cells compared with the neighboring wild-type cells (supplementary material Fig. S2). In summary, these results demonstrate that overexpression of miR-275 or miR-306, but not miR-7 or miR-317, is sufficient to repress Bam protein accumulation.

Consistent with the function of miR-275 in downregulating Bam protein in spermatocytes, in situ hybridization using miR-275 probe showed enriched miR-275 expression in early spermatocytes (supplementary material Fig. S3A). Although miR-306 expression pattern shown by in situ was broader, it was also detectable in spermatocytes (supplementary material Fig. S3B). These data suggest that both miR-275 and miR-306 could cooperate in downregulating Bam protein in spermatocytes.

It has been shown that a tandem-repeat sequence of a microRNA binding site called sponge (SP) sequence can downregulate microRNAs (Ebert et al., 2007; Loya et al., 2009). In order to investigate whether knockdown of miR-275 and/or miR-306 in spermatocytes stabilizes Bam, we expressed UAS-GFP-miR275SP or UAS-GFP-miR306SP using the bam-Gal4; UAS-Gal4 drivers to achieve a high expression in 4-16 spermatogonia and all spermatocytes (supplementary material Fig. S4A). To enhance SP knockdown efficiency, we removed one copy of the dicer-1 (dcr-1) gene using a null allele dcr-1^Q1147X to further suppress microRNA biogenesis (Lee et al., 2004). However, neither the expression of miR-275SP or miR-306SP alone (supplementary material Fig. S4B,C) nor the co-expression of miR-275SP and miR-306SP (supplementary material Fig. S4D) stabilized Bam in spermatocytes (supplementary material Fig. S4B,C). It is possible that expression of these SP sequences was insufficient to knock down robust expression of microRNAs in spermatocytes. Alternatively, it is possible that the Gal4; UAS system does not work effectively in spermatocytes and the sustained GFP reporter in spermatocytes (supplementary material Fig. S4A) is due to perdurance of the GFP protein. Last, it is possible that some RNA-binding proteins act through the same 3′UTR sequences to regulate Bam accumulation, a mechanism that we did not examine in this study.

The failure to downregulate Bam in spermatocytes could have biological consequences. To understand the scope of such consequences, we performed a fertility assay of bam-HA bam3′ and bam-HA tub3′ transgenic males. Interestingly, we found a significant increase (P<0.05) of male sterility in bam-HA tub3′ transgenic males, in which Bam persisted in spermatocytes. Approximately 13% of bam-HA tub3′ transgenic males were sterile (n=149), whereas only 3% of bam-HA bam3′ transgenic males were sterile (n=99). To further investigate the defects leading to sterility of bam-HA tub3′ transgenic males, we dissected seminal vesicles, where mature sperm are stored, from sterile males and examined sperm motility using phase-contrast microscopy. Although all seminal vesicles from wild-type males contained motile sperm (n=30), 90% of the seminal vesicles (n=20) from bam-HA tub3′ sterile males had no motile sperm.

During Drosophila spermiogenesis, interconnected spermatids descending from germline syncytia undergo a process called individualization (Fuller, 1993; Lindsley and Tokuyasa, 1980). During this process, individualization complexes containing 64 actin cones move along the sperm axoneme to generate individual sperm (Fabrizio et al., 1998; Noguchi and Miller, 2003). Using phalloidin staining, we quantified these actin cone bundles or individualization complexes. On average, 22.0 individualization complexes per testis were detected in wild-type testes (Fig. 6A,C), whereas only 6.6 individualization complexes per testis were detected in testes from bam-HA tub3′ sterile males (Fig. 6B,C). These results demonstrate that failure in downregulation of Bam in spermatocytes leads to spermiogenesis defects and increased male sterility.

In the male germline, it has been shown that ectopic expression of Bam in male GSCs induces their loss, probably by precocious differentiation (Schulz et al., 2004; Sheng et al., 2009). Our data showed that Bam expression must be repressed in spermatocytes for proper terminal differentiation during spermiogenesis (Fig. 6D). Therefore, although Bam functions as a differentiation gene for GSCs, it must be downregulated in spermatocytes for proper terminal differentiation. These results suggest that Bam plays distinct roles at different stages of spermatogenesis and that the microRNA-mediated post-transcriptional regulation of Bam is specific for differentiating spermatocytes, but not for GSCs. Previous studies have focused on the molecular mechanisms that prevent ectopic Bam expression in GSCs. Our data reveal a novel regulatory mechanism that prevents Bam accumulation in differentiating spermatocytes, which is crucial for their proper terminal differentiation into functional gametes. Therefore, our discoveries highlight the specificity of developmentally programmed cellular differentiation pathways.

In the male GSC lineage, Bam is required for the transition from proliferative stage to differentiation stage. Previously, it was shown that accumulation of Bam to the threshold level is important to induce the transition from spermatogonia to spermatocytes (Insco et al., 2009). Indeed, either bamHA bam3′ or bamHA tub3′ transgene, on top of the endogenous Bam, induced early transition from spermatogonia to spermatocytes, which resulted in increased testes containing at least one eight-cell spermatocyte cyst. However, in the previous study (Insco et al., 2009), a transgene that is almost identical to the bam HA bam3′ we used in this report did not lead to premature spermatogonia-to-spermatocyte transition. This discrepancy could have resulted from the different expression level of transgenes. In our assay, site-directed incorporation was used for all transgenes. The insertion site attP40 has been optimized for high expression (Markstein et al., 2008), whereas in the previous assay, P-element-mediated random incorporation was used, which may not lead to high expression of transgenes.

However, the observation that bam mRNA is present while Bam is absent in early spermatocytes provokes another possibility: bam mRNA may have novel functions in early spermatocytes that are independent of Bam. Such functions remain unexplored in bam mutant testes because germ cells are arrested at the spermatogonial stage and cannot differentiate into early spermatocytes. Further studies will be carried out to knock down bam mRNA specifically in early spermatocytes and address this possibility. Alternatively, the previous study (Insco et al., 2009) directly quantified the percentage of eight-cell spermatocyte cysts, whereas we quantified the percentage of testes containing at least one eight-cell spermatocyte cyst (see Materials and methods). It is possible that our method produces a higher ratio compared with the previous study.

Bam is an essential differentiation factor in multiple stem cell lineages, including male and female GSC lineages (Gönczy et al., 1997; McKearin and Spradling, 1990). In both male and female GSCs, TGFβ signaling activated by ligands emanated from somatic cells in the niche represses bam transcription (Kawase et al., 2004; Schulz et al., 2004; Shivdasani and Ingham, 2003). In both male and female gonads, ectopic expression of Bam in GSCs induces differentiation and results in loss of GSCs (Ohlstein and McKearin, 1997; Sheng et al., 2009). Despite a high degree of similarity between male and female GSC lineages and Bam expression pattern, our data demonstrate that the microRNA-controlled post-transcriptional regulation of Bam is specific for male, but not for female, gametogenesis. Therefore, our data provide a new line of evidence showing distinct molecular mechanisms employed in male versus female germline stem cell differentiation pathways.

However, in the lymph gland, Bam maintains hematopoietic progenitors at an undifferentiated state and bam mutation results in a reduction of hematopoietic progenitors, but an increase in differentiated hemocytes (Tokusumi et al., 2011). Therefore, in different tissues, Bam acts differently to either block or induce differentiation.

Previously, it has been suggested that miR-7 binds to bam 3′UTR and that overexpression of miR-7 in early germ cells increases the number of spermatogonial cysts (Pek et al., 2009). However, overexpression of miR-7 did not affect Bam expression in our analysis. We speculate two possible explanations for this discrepancy. First, in our study, miR-7 might not be expressed at a high enough level to change Bam expression. Second, we used a different reporter in our assay. We used the bam-HA transgene as the reporter, which has endogenous promoter, bam-coding sequence and bam 3′UTR in germ cells, whereas the previous study used the reporter with tubulin promoter, GFP coding sequence and bam 3′UTR in somatic cells. In addition, as readout for decreased Bam protein levels, we used the percentage of testes containing at least one 32-spermatocyte cyst, which reflects Bam protein levels precisely (Insco et al., 2009); the previous study used nuclear morphology (Pek et al., 2009). In hematopoietic progenitors, Bam and miR-7 function cooperatively, but bam is not a target of miR-7 (Tokusumi et al., 2011).

Tight regulation of adult stem cell differentiation in various lineages is vital because disruption of such regulation can lead to tumorigenesis (Clarke and Fuller, 2006). MicroRNAs have been reported to play an essential role in regulating stem cell self-renewal and differentiation (Gangaraju and Lin, 2009). It has also been shown that microRNA expression is misregulated in many cancers (Lu et al., 2005; Volinia et al., 2006). Although it has been proposed that microRNAs contribute to stem cell maintenance and proper differentiation, most of the evidence in support of this has been based on the phenotypes caused by mutating the enzymes responsible for the biogenesis or stabilization of microRNAs. For example, mutations of the microRNA biogenesis components dcr-1, loquacious, argonaut 1 and mei-P26 lead to female GSC loss (Förstemann et al., 2005; Jin and Xie, 2007; Li et al., 2012; Park et al., 2007; Yang et al., 2007). It is much less understood how individual microRNAs regulate their endogenous target genes during stem cell differentiation programs in vivo. Our data provide a key insight into how microRNAs regulate Drosophila male germline stem cell lineage. Delay of the proliferation-to-differentiation transition upon overexpression of miR-275 or miR-306 in testes indicates that microRNA expression also needs to be tightly controlled. Interestingly, either miR-275 or miR-306 can affect the Bam expression levels. Therefore, a single microRNA can have multiple targets; in addition, a single gene can be regulated by multiple microRNAs. Tight regulation of gene expression can be partially achieved by spatial- and temporal-specific expression of microRNAs.

We thank Drs Dennis McKearin and Jean Maines for the plasmid containing bam-HA transgene. We also thank Dr Mark Van Doren and Chen lab members for their critical reading and insightful suggestions.