Gilgamesh is required for the maintenance of germline stem cells in Drosophila testis

Emerging evidence supports that stem cells are regulated by both intrinsic and extrinsic mechanisms. However, factors that determine the fate of stem cells remain incompletely understood. The Drosophila testis provides an exclusive powerful model in searching for potential important regulatory factors and their underlying mechanisms for controlling the fate of germline stem cells (GSCs). In this study, we have found that Drosophila gilgamesh (gish), which encodes a homologue of human CK1-γ (casein kinase 1-gamma), is required intrinsically for GSC maintenance. Our genetic analyses indicate gish is not required for Dpp/Gbb signaling silencing of bam and is dispensable for Dpp/Gbb signaling-dependent Dad expression. Finally, we show that overexpression of gish fail to dramatically increase the number of GSCs. These findings demonstrate that gish controls the fate of GSCs in Drosophila testis by a novel Dpp/Gbb signaling-independent pathway.


Results
Identification of gish mutants with defects in GSC maintenance. To identify novel genes that regulate the self-renewal or differentiation of GSCs in Drosophila testis, we conducted a screen for male sterile mutants with P-element insertion, available from the Bloomington Stock Center. We isolated a line with a P-element insertion in the third chromosome 15 , P{PZ}gish 04895 , which resulted in small testis and reduction in germ-cell number in homozygous mutant males. To thoroughly analyze the behavior of germ cells in gish 04895 mutant, we used anti-Fas III and anti-Vasa antibodies to visualize hub and germ cells 17,18 , respectively (Fig. 1b). Hub is a cluster of somatic cells found at the tip of the adult testis which could be characterized by anti-Fas III antibody staining in Drosophila. As shown in Fig. 1b, we visualized GSCs and germ cells with an anti-Hts antibody 19 . A spherical fusome (also called the "spectrosome") is the marker of GSCs and GB. GB undergoes 4 rounds of cell division to produce a 16-cell germline cluster, in which branched fusomes are visualized by the anti-Hts antibody (Fig. 1a,b).
In the tip of wild-type testis, 6-10 GSCs that were directly adjacent to the hub were recognized by an anti-Vasa antibody (Fig. 1b) 17 . In contrast, in the 10-day-old testes from gish 04895 homozygous males, about 30-40% of mutant testes (n > 100) contained 3-4 GSCs attached to the hub. This finding suggests that gish may plays a key role in the maintenance of GSCs. To explore whether gish is involved in the GSC maintenance in other relevant genetic background, we next performed phenotypic analyses with gish removal in three allelic combinations (Table 1). Using the methods described in the previous study 17 , we counted the number of GSCs in gish allelic combinations at days 1, 10 and 20 after eclosion. Compared to wild-type and gish/+ heterozygotes, the newly Testes stained with anti-Fas III antibody to label the hubs (red, indicated by asterisks), anti-Hts antibody to label the fusomes (red), and anti-Vasa antibody to label germ cells (green). GSCs were highlighted by white dots. (b) In wild-type testis, seven GSCs (white dot) directly contact the hub (asterisk). (c) gish 04895 mutant testis from 1-day-old fly showing that six GSCs contact the hub. (d) gish 04895 /gish MI08417 testis from 10-day-old fly showing that five GSCs remains adjacent to the hub. (e) gish KG03891 /gish MI08417 testis from 20-day-old fly showing three remaining GSCs. (f) gish 04895 /gish MI08417 testis from 20-day-old fly. Only two GSCs remains. (g) Quantitative PCR analyses of gish mRNA levels in testes between gish mutant and wild-type. (h) The transgene P{gishP-gishF} rescued the gish 04895 /gish MI08417 male testis to normal, with eight GSCs close to the hub. Scale bars: 10 μm.
eclosed gish mutants (gish 04895 /gish KG03891 , gish 04895 /gish MI08417 , gish KG03891 /gish MI08417 ) contained an average of 6.7, 6.5 and 6.3 GSCs/testis ( Fig. 1c and Table 1). When these three gish mutants were cultured at room temperature for 10 days, the testes had an average of 5.4, 5.7 and 5.8 GSCs/testis respectively ( Fig. 1d and Table 1), whereas the testes from wild-type contained a normal average of 7.7 GSCs/testis. At day 20, the average number of GSCs was dramatically reduced to 3.9, 3.7 and 3.5 GSCs/testis respectively (Fig. 1e,f and Table 1), compared to the average number of wild-type maintained at the normal level (6.0 GSCs/testis) ( Table 1). The testes from gish 04895 /Df and gish KG03891 /Df exhibited a similar loss of GSCs with ageing (Table 1). These statistical data indicate that the deficiency of gish leads to a progressive loss of GSCs with ageing.
To test whether the GSC loss phenotype is due to a reduced gish expression level in gish mutant testis, we performed real-time quantitative polymerase chain reaction (qPCR) experiments to compare the mRNA level between the wild-type and mutant fly testis 20 . The gish gene totally has thirteen predicted mRNA splicing variants (A, B, C, D, E, F, G, H, I, J, K, L and M) derived from the Drosophila database (www.flybase.org), but their corresponding encoded proteins share a conserved kinase domain. Based on this, the qPCR primers were designed (see the materials and methods) to target the conserved cDNA region using the online Primer 3.0 version software to detect the whole mRNA expression level of gish. We extracted total RNA from Drosophila testes, performed reverse-transcription (RT), and conducted qPCR experiments to measure the whole gish mRNA level with the rp49 gene as reference 21 . Compared to wild-type, the total gish mRNA expression levels in gish mutant testes (gish 04895 /gish KG03891 , gish KG03891 /gish MI08417 and gish 04895 /gish MI08417 ) were severely reduced (Fig. 1g). These results strongly suggest that Gish is reduced in these gish mutants' testes, and also imply that the Gish protein is responsible for the loss of GSCs phenotype in gish mutant flies.
To further confirm the role of gish in GSC maintenance, we next performed a gish rescue assay using gish cDNA. The gish gene contains numerous splicing variants as mentioned above. To explore which isoforms are expressed in Drosophila testis, we designed a dozen of primer pairs (Supplementary Table S1), each for each known gish transcript, and performed RT-PCR analysis 22 using total RNAs isolated from wild-type fly testis. We detected transcripts gishF and B in testis ( Supplementary Fig. S1). Based on these findings, we generated a transgene of P{gishP-gishF}, in which the gishF cDNA was placed under the control of a 5.0 kb gish promoter. We found that the GSC loss phenotypes in three gish allelic mutants were fully rescued by the transgenic line of P{gishP-gishF} ( Fig. 1h and Supplementary Table S2). Taken together, our findings demonstrate that the gish gene plays an essential role in GSC maintenance.
Previous study has reported that gish mutant germ cells contain abnormal actin cones in individualizing spermatid cysts 15 , this result implies that gish possibly affects the F-actin-mediated cell adhesion junctions. To explore whether the gish mutant GSCs lose adhesion to the hub, we visualized F-actin with phalloidin 15 and labeled GSCs with anti-Vasa antibody 17 . We found that there was no difference in F-actin-based GSCs adhesion to hub cells between wild-type control and gish KG03891 /gish MI08417 mutant testes (n > 80) at day 14 after eclosion ( Supplementary Fig. S2), just as the wild-type, GSCs from gish mutant testes contacted directly to hub cells. Similar phenotype was founded in gish 04895 /gish MI08417 mutant testes (n > 70). These results manifest that the mutation of gish gene doesn't affect cell-cell (GSC and hub cell) adhesions in fly testes, suggesting some other mechanisms maybe responsible for the GSCs loss phenotype.
To address whether the loss of GSCs in gish mutants was due to premature differentiation or cell death of GSCs, we measured the rate of apoptosis in the gish mutant GSCs by TUNEL assays 23 . We found that there was no difference in apoptosis between wild-type control and gish mutant GSCs (Supplementary Table S3 and Genotype The average number of GSCs in fly testis with the elapse of days (Mean ± SD)
The gish gene is intrinsically required for GSC maintenance. Previous studies have shown that the maintenance of GSCs is regulated through intrinsic and extrinsic signaling pathways in testis 24,25 . To explore the role of gish in GSC maintenance, we generated a transcriptional reporter P{gishP-GFP}, in which the gfp expression pattern represents that of gish gene 26 . As shown by immunofluorescent staining assays ( Fig. 2a and Supplementary Fig. S4), GFP was ubiquitously expressed in all cell types including somatic cells (e.g. hub) and germline cells (e.g. GSCs and GBs) in transgenic fly testes (n > 100), suggesting a broad transcription activity of the gish promoter.
To determine whether gish acts as an intrinsic or extrinsic regulator, we then generated gish mutant GSC clones using the FLP/FRT-mediated mitotic recombination technique [27][28][29] . The gish mutant GSCs were negatively-marked after several days of heat-shock treatments. We counted and compared the loss rate of marked GSC clones, as described previously 17,23 , between the FRT control (hs-flp; FRT82B/FRT82B, ubi-gfp) and the gish mutant genotype (hs-flp; FRT82B, gish/FRT82B, ubi-gfp), at days 2, 10 and 20 after heat-shock treatments (AHST). As shown in Fig. 2, the rates of marked GSC clones from FRT control decreased weakly from the initial of 72.0% (n = 254) at day 2 to the last point of 59.9% (n = 197) at day 20 AHST (Fig. 2b,c and k). Only 16.8% of the marked GSCs were lost during the 20-day AHST period. By contrast, under the same experimental conditions, the initial marked gish 04895 and gish MI08417 mutant GSCs clones were 66.6% (n = 213) and 66.4% (n = 188) respectively at day 2 AHST, whereas they reduced to the rates of 21.2% (n = 189) and 24.4% (n = 279) respectively at day 20 AHST (Fig. 2d-f and k). These results revealed that 68.2% and 63.3% of marked gish 04895 and gish MI08417 mutant GSCs clones were lost during the 20-day AHST. Taken together, these findings suggest that gish plays an essential role for GSC maintenance through an intrinsic mechanism. To further confirm this point, we generated a transgene, P{nosP-gishF}, in which the gishF coding sequence was placed under the control of the promoter of the gene nanos that possesses a high expression level in germline cells. We found that the GSCs loss phenotype was fully rescued in gish mutant flies carrying the P{nosP-gishF} transgene (Fig. 2g,h and Supplementary Table S4). This result supports that Gish functions as an intrinsic factor.
To further test whether gish also plays an extrinsic role in GSC self-renewal, we generated a new transgenic fly strain carrying P{UASp-gishF}, in which the gishF coding sequence is under the control of the UASp promoter 30 . Using the Gal4-UAS system, we expressed the Gish protein in somatic hub cells and somatic stem cells by the c587-gal4-driven UASp-gishF expression 17 . We found that the GSCs loss phenotype was not rescued in gish mutant testes carrying the c587-gal4 and UASp-gishF transgenes ( Fig. 2i and Supplementary Table S5). We then expressed gish specifically in germ cells that carried transgenes of P{UASp-gish} and P{nosP-gal4}, in which a germline-specific nanos-gal4 driver can express the target gene under the control of UASp promoter 31 . We found that the gish phenotype was fully rescued (Fig. 2j and Supplementary Table S6). This result is consistent with the above observation on the rescuing effect of P{nosP-gishF}. Taken together, these data strongly indicate that gish is an intrinsic factor that regulates GSC self-renewal.
gish is not required for Dpp/Gbb signaling silencing of bam. A Previous study has shown that two BMP members, Dpp and Gbb function cooperatively to maintain GSCs in Drosophila testis by repressing bam transcription in GSCs 17 . To explore whether gish is involved in Dpp/Gbb-dependent bam silencing, we examined the bam expression pattern in gish mutant testes through the expression of bam transcriptional reporter P{bamP-GFP} 32 . As shown in Fig. 3, the germ cells in testes from 5-day-old flies after eclosion were labelled with anti-GFP antibody and DAPI staining. We found that 88.5% of GSCs and 83% of GBs (n = 65 testes) from the male wild-type flies carrying P{bamP-GFP} reporter exhibited a completely negative GFP pattern (Fig. 3a). Similar phenotypes were observed in gish mutant testes, 89% of GSCs and 86.5% of GBs (n = 70 testes) showed to be negative for GFP (Fig. 3b). These data strongly suggest that gish is not required for bam silencing.

Gish is dispensable for dpp/gbb signaling-dependent Dad expression. It has been reported that
dpp signaling is necessary for Dad expression, whereas Dad negatively modulates dpp signaling, forming a negative-feedback loop in Drosophila wing development 33,34 . Another study shows that Dad is a gbb-responsive gene that negatively regulates gbb signaling in GSCs and GBs in Drosophila testis 17 . To test whether mutation of gish affects Dpp/Gbb signaling-dependent Dad expression in GSCs, we examined the expression of Dad transcriptional reporter P{DadP-GFP} in the gish mutant background. After gish mutant males were cultured at 25 °C for five days, 93% of GSCs and 94.5% of GBs (n = 65 testes) of wild-type flies carrying P{DadP-GFP} reporter showed positive for GFP (Fig. 3c). Similarly, 95% of GSCs and 90% of GBs (n = 68 testes) from gish mutant testes exhibited a GFP-positive pattern (Fig. 3d). These results convincingly suggest that gish is dispensable for Dad expression.
Ectopic gish expression had no effects on the number of GSC. Loss of function of gish contributed to testis GSC loss without involvement of apoptosis, it is likely that overexpressed gish in testis may delay GSCs/GBs differentiation. To test this possibility, we isolated GSCs from somatic cells and differentiated germ cells by using anti-Hts, anti-fas III and anti-Vasa antibodies to visualize fusomes, hub cells and germ cells, respectively 17,18 . The spectrosome, a spherical fusome, is the marker of GSCs and their early progeny (also known as GB). When a late-stage GB divides to 2-cell, the fusome is visualized as a short-bar connecting two GB cells. GB undergoes 2, 3 and 4 rounds of synchronous cell division to produce a 4, 8 and 16-cell germline cluster, in which branched fusomes were visualized by anti-Hts antibody (Fig. 1a). We counted the number of germ cells carrying spectrosomes in testes from wild-type, and P{nosP-gishF} transgenic flies 7 days after eclosion. We observed the averages of 11.3 spectrosome-contained GSC/GBs (n = 67 testes) per testis in wild-type and 12.5 spectrosome-contained Scientific RepoRts | 7: 5737 | DOI:10.1038/s41598-017-05975-w  Table 2 and Fig. 4a and b). These results suggest that there is no significant difference in the number of spectrosome-contained germ cells among wild-type and gish-overexpression transgenic flies.
To substantiate this result, we generated transgenic flies carrying P{hs-gishF}, in which a gishF cDNA was placed under the control the heat-shock promoter 35 . We overexpressed gishF in testes by applying heat-shock treatment at 37 °C for 1.5 hours three times each day, counted the number of germ cells carrying spectromes in testes 7 days after heat-shock treatment with P{hs-gishF} flies cultured at 25 °C as control. We found an average of 11.6 GSC/GBs carrying spectrosomes per testis (n = 69 testes) in control flies (Table 2 and Fig. 4c). In contrast, in testes with ectopic gishF expression, the number of the germ cells containing spectrosomes was increased to an average of 14.8 cells per testis (n = 72 testes) ( Table 2 and Fig. 4d). These results also show that there is no obvious increase in the numbers of spectrosome-containing germ cells after ectopic gish expression in P{hs-gishF} transgenic flies. Taken together, these data imply that increased Gish activity is not sufficient to block GSC/GB differentiation.

Discussion
Past research has demonstrated that male flies with homozygous gish of a P element insertion (P{ry +t7.2 = PZ}gish 04895 ) exhibited sterile and defective spermatid individualization in Drosophila testes 13 . Our previous observation found that the testis became very small in size in gish 04895 mutant fly. These thinned testes prompted us to explore whether gish is involved in the maintenance of germline stem cells in Drosophila testis. Combining germline clonal analysis and rescue tests, we showed that gish plays an intrinsic role in GSC self-renewal, suggesting that gish is necessary to regulate GSCs' fate. In addition, we found that ectopic gish expression only slightly increased the number of GSC-like cells. These results suggest that overexpression of gish is not sufficient to repress GSCs/GBs differentiation. Previous studies demonstrated that mutation in gish gene led to abnormal nuclei and altered structure of actin cones in the individualizing spermatid cyst 15,36 , but no further information is available during the early process of GSCs maintenance or switch to differentiation. Our results suggest that gish plays a novel role as a GSC intrinsic maintenance factor, but has no roles in the differentiation of GSC/GB in Drosophila testis. Previous research has manifested that Bmp signals from somatic cells, Dpp and Gbb, are essential for the maintenance of GSCs in the Drosophila testis 17 . Both Gbb and Dpp function as short-range signals in the tip of the Drosophila testis, and their signaling activities are restricted to GSCs and GBs 17,33,34 . Interestingly, Dad is expressed in GSCs and GBs, whereas bam is not expressed in either kind of cell. Both of Dpp and Gbb are essential for maintenance of GSC number 17 . Thus, we checked the bam and Dad expression pattern using bam-GFP and Dad-GFP transgene. The results show that the mutation in gish has no effect on the expression pattern of bam and Dad in GSCs and GBs in Drosophila testes. These data suggest that gish functions downstream of or parallel to Dad/bam, and is independent of Gbb/Dpp-bam signaling pathway.

Materials and Methods
Drosophila stocks. Oregon-R was used as a wild-type strain. The w 1118 strain was used as the host for all P-element-mediated transformations 37 . The following strains were also used for experimentation: (1) gish 04895 , gish MI08417 and gish KG03891 alleles (Bloomington Stock Center); (2) P{bamP-GFP} 38 ; (3) P{nosP-gal4}, which was described previously 31 . P{dadP-GFP}, neoFRT82B/TM3 and hs-FLP; FRT82B, Ubi-GFP/TM 3 was a generous gift from Dr. Dahua Chen. Fly stocks used in this study were maintained at room temperature on a standard medium.
Histochemistry and microscopy. Testes were prepared for immunostaining as previously described 17,23 .
Generation and Analysis of GSC Clones. Mutant GSC Clones were generated by FLP/FRT-mediated mitotic recombination, as described previously 28,39 . To generate stocks for stem cell clonal analyses, males of hs-FLP; FRT82B,Ubi-GFP/FRT82B,gish 04895 and hs-FLP; FRT82B,Ubi-GFP/FRT82B,gish MI08417 genotypes (hs-FLP; FRT82B, Ubi-GFP/FRT82B as the wild-type control) were produced by standard genetic crosses. 2-day-old adult males were heat-shocked for 60 minutes at 37 °C three times per day. Five days after heat-shock treatment, testes were dissected for antibody staining at days 2, 10, 20 after last heat-shock treatment. GSC clones were identified and quantified by the lack of GFP expression, as well as their attachment position to the hub cells.
Detecting gene expression in Drosophila testis using RT-PCR and qPCR. Total RNA was extracted from wild-type fly testes using Trizol reagent (Sangon), then the amount of 1 µg was incubated with PrimeScript RTase(50 U/µl) to transcribe cDNA, which was used as the template in PCR reactions, according to the manufactures' protocol (PrimeScript High Fidelity RT-PCR Kit, Takara). The PCR primers were designed to explore the different gish mRNA splicing variants (S1 Table). Total RNAs of fly testes were independently isolated from each phenotype (wild-type and gish mutant) using the Trizol reagent (Sangon) and reverse transcribed into cDNA according to the manufactures' protocol (PrimeScript RT reagent Kit with gDNA Eraser, Takara). For each independent cDNA sample, quantitative PCR was run on CFX96 Touch (BioRad) to measure total gish mRNAs with rp49 as reference according to the manufactures' protocol (SYBR Premix EX Taq TM II qPCR Kit, Takara). The following primers were used in this study: gish, 5′-GCCGGTGGTAAAAGCTCAAG-3′ (sense) and 5′-CGCCAAAATTACCACAGCCA-3′ (antisense); rp49, 5′-CACTTCATCCGCCACCAGTC-3′ (sense) and 5′-CGCTTGTTCG ATCCGTAACC-3′ (antisense).