identification of seven genes essential for male fertility through a

OR I G INA L ART I C L E

Identification of seven genes essential for male fertilitythrough a genome-wide association study ofnon-obstructive azoospermia and RNA interference-mediated large-scale functional screening in DrosophilaJun Yu1,2,†, HaoWu1,2,†, YangWen1,4,†, Yujuan Liu1,2,†, Tao Zhou1,2, Bixian Ni1,4,Yuan Lin1,4, Jing Dong1,4, Zuomin Zhou1,2, Zhibin Hu1,4,*, Xuejiang Guo1,2,*,Jiahao Sha1,2,*, and Chao Tong3,*1State Key Laboratory of Reproductive Medicine, 2Department of Histology and Embryology, Nanjing MedicalUniversity, Nanjing 210029, China, 3Life Sciences Institute and Innovation Center for Cell Biology, ZhejiangUniversity, Hangzhou 310058, China, and 4Department of Epidemiology and Biostatistics, School of Public Health,Nanjing Medical University, Nanjing 211166, China

*To whom correspondence should be addressed at: Life Sciences Institute, Zhejiang University, 866 Yuhangtang Road, Hangzhou, Zhejiang 310058, China.Tel: +86 57188981370; Fax: +86 57188981337; Email: [email protected] (C.T.); State Key Laboratory of Reproductive Medicine, Nanjing Medical University,140 Hanzhong Road, Nanjing, Jiangsu 210029, China. Tel: +86 2586862038; Fax: +86 2586862908; Email: [email protected] (J.S.); [email protected](X.G.); Tel: +86 2586868440; Fax: +86 2586868439; E-mail: [email protected] (Z.H.)

AbstractNon-obstructive azoospermia (NOA) is a complex and severe condition whose etiology remains largely unknown. In a genome-wide association study (GWAS) of NOA in Chinese men, few loci reached genome-wide significance, although this might be aresult of genetic heterogeneity. Single nucleotide polymorphisms (SNPs) without genome-wide significance may also indicategenes that are essential for fertility, and multiple stage validation can lead to false-negative results. To perform large-scalefunctional screening of the genes surrounding these SNPs, we used in vivo RNA interference (RNAi) in Drosophila, which hasa short maturation cycle and is suitable for high-throughput analysis. The analysis found that 7 (31.8%) of the 22 analyzedorthologous Drosophila genes were essential for male fertility. These genes corresponded to nine loci. Of these genes, leukocyte-antigen-related-like (Lar) is primarily required in germ cells to sustain spermatogenesis, whereas CG12404, doublesex-Mab-related11E (dmrt11E),CG6769, estrogen-related receptor (ERR) and sulfateless (sfl) function in somatic cells. Interestingly, ERR and sfl are alsorequired for testis morphogenesis. Our study thus demonstrates that SNPs without genome-wide significance in GWAS mayalso provide clues to disease-related genes and therefore warrant functional analysis.

Introduction

More than 10% of couples of childbearing age have fertility pro-blems. Male factor infertility accounts for approximately half of

these cases (1). The causes of male infertility are complex anddiverse, but non-obstructive azoospermia (NOA), themost severeform of male infertility, occurs in 1% of all adult men (1). Thesepatients have no sperm in their semen. In the past few years,

† These authors contributed equally to this work.Received: July 19, 2014. Revised: October 21, 2014. Accepted: October 27, 2014.

© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Human Molecular Genetics, 2015, Vol. 24, No. 5 1493–1503

doi: 10.1093/hmg/ddu557Advance Access Publication Date: 30 October 2014Original Article

1493

Downloaded from https://academic.oup.com/hmg/article-abstract/24/5/1493/796023by gueston 02 April 2018

http://www.oxfordjournals.org

some genes associated with male infertility have been identified(2,3), yet we still know little about NOA.

Since the first genome-wide association study (GWAS) wassuccessfully performed in 2005 (4), thousands of single nucleo-tide polymorphisms (SNPs) have been found to be associatedwith common diseases (5). GWAS help to connect SNPs with cer-tain diseases and can be used to improve risk assessment andaccelerate drug development. However, most GWAS have lackedsystematically functional verification, limiting their role in guid-ing our understanding of humandiseases. In addition, huge sam-ple sizes and strict quality control are needed to uncover loci withsignificant genome-wide associations (P-value < 5 × 10−8) (6,7).Large numbers of SNPs with weak associations are filtered outowing to a potential limiting factor in the statistical power to de-tect weak genetic loci (8). Furthermore, GWAS do not attempt toidentify functional SNPs but rather ‘tag’ the approximate locationof disease variants, and they can identify disease-associated al-leles only when such alleles have strong linkage disequilibriumwith tag-SNPs (tSNPs). Such associationswill beweak for raredis-ease-causing variants (9). Rare variants tend to show greater geo-graphic clustering than common ones and can display spatialstructures in populations that are not well controlled in commonvariant analyses (10). Therefore, tSNPs with lower associationscoresmay be excluded bymultiple validation analysis, resultingin false-negative results.

In our previouswork, we performed a three-stage analysis of aNOA GWAS in Han Chinese men (a total of 2927 individuals withNOA and 5734 control cases) (11). After the combined analysis,only three SNPs (rs12097821, rs2477686 and rs10842262) confer-ring a significant genome-wide risk of NOA (P-value < 5 × 10−8)were identified. Pex10, a gene adjacent to the risk locusrs2477686, was reported to be essential for male fertility in Dros-ophila (12), indicating the effectiveness of GWAS at identifyinggenes required for male fertility. Additionally, our extended val-idation sample size of 2806 cases and 4334 controls revealedthree additional NOA susceptibility loci with genome-wide sig-nificant associations as well as one locus approaching genome-wide significance (rs3000811, upstream of CDC42 binding proteinkinase alpha; CDC42BPA). We also found that gek, an ortholog ofCDC42BPA, is essential for male fertility in Drosophila (13).

Because the identified susceptibility loci far from fully explainthe complexity of human NOA, we hypothesize that SNPs with-out genome-wide association (or without the support ofmultiplevalidation analysis) may also be important indicators of causalgenes for NOA. To test this hypothesis, we developed a strategyusing a Drosophila in vivo RNA interference (RNAi) knockdownsystem to systematically explore and identify functional sperm-atogenesis genes from tSNPs identified by GWAS.

Drosophila is an excellent model organism in which to screengenes essential for male fertility (14). Drosophila has a shortlife cycle, and genome-wide gene-knockdown resources areavailable, which make high-throughput functional screeningpossible. AdultDrosophila testes contain spermatogonia, sperma-tocytes, spermatids and sperm, the same populations of germcells that exist in the human testis. At the apical tip of the Dros-ophila testis, germline stem cells (GSCs) divide to generate twocells, a new stem cell and a gonialblast that proliferates and dif-ferentiates into spermatocytes (15). The cyst cells provide an en-vironment for germ cell proliferation, growth and differentiation,similar to the role of Sertoli cells in the human testis (16). Manyfertility-related genes are conserved between Drosophila andmammals (17), and many mutants of homologous human andDrosophila genes have similar testicular phenotypes (14,18).Using the Drosophila in vivo RNAi model, here we carried out a

systematic analysis of genes surrounding the loci associatedwith NOA and identified seven genes (7/22, 31.8%) essential formale fertility.

ResultsA strategy for male fertility-essential genes by GWAS ofNOA followed by RNAi-mediated functional screening inDrosophila

In our previous NOA GWAS screen, 109 tSNPs were found to beassociated with NOA (P-value < 10−5). After several rounds ofvalidation, we found strong evidence of genome-wide sig-nificance (P-value < 5 × 10−8) for NOA susceptibility for six SNPs,as well as one locus that approached genome-wide significance.Although the other SNPs did not reach genome-wide significance(P-value < 5 × 10−8) by multiple validation analysis, they may stillindicate surrounding genes essential for fertility that confer arisk for NOA.

A recent study showed that the majority (∼93%) of disease-and trait-associated SNPs identified by GWAS lie within non-coding sequences. Approximately 32% of GWAS SNPs are locatedin deoxyribonuclease I hypersensitive sites (DHSs), markers ofregulatory DNA that can regulate genes within 100 kb (19). Tocover more regulated genes, we considered genes flanking thetSNPs within 200 kb. For the 103 tSNPs that did not reach gen-ome-wide significance after multiple validations but had a P-value < 10−5, genes were classified into two classes: Class I,genes with the tSNP located inside the gene body; Class II, adja-cent genes located within the 200 kb flanking the tSNPs. In total,142 candidate genes were found around the tSNPs in the humangenome. For Drosophila ortholog annotation, only orthologs(homology type: one to one and one to many) with sequenceidentity >20% were considered for functional studies. The ana-lysis found 38 orthologous Drosophila genes through the Ensembldatabase (http://www.ensembl.org/), corresponding to 37 humangenes and 38 susceptible tSNPs. Of these genes, NfI and sina areessential for male fertility according to Flybase (http://flybase.org/) and the MGI database (http://www.informatics.jax.org/).We have previously shown that gek is essential for fertility (13).Additionally, mice or Drosophila carrying mutations in six othergenes (Nak, Fak56D, Oseg5, CG9220, CG3925 and IA-2) are knownto be fertile. UAS-RNAi lines for another seven orthologousDrosophila genes (corresponding to four human genes and fourcandidate tSNPs) are not available. These genes were excluded,and the remaining 22 genes were subjected to functional studies(Fig. 1 and Supplementary Material, Table S1).

The Drosophila UAS-Gal4 system-based RNAi silencing meth-odhas previously beenused for fertility analysis (20). Because dif-ferent Gal4 drivers have different expression levels and patterns,seven different Gal4 lines were used to drive UAS-RNAi expres-sion in the fly testis (see Fig. 2B). Nos-Gal4 has germ cell-specificexpression, primarily in early stage germ cells (21,22). Bag ofmarbles-Gal4 (bam-Gal4) expression is restricted to 2- to 16-cellspermatogonia (22,23). C587-Gal4 is expressed in cyst cells (24).C729-Gal4 is expressed in cyst cells, the testis sheath and themale accessory gland (13). Esg-Gal4 is expressed in the somaticcells of the developing gonad (25), and upd-Gal4 is expressed inthe hub cells (22). We also used the weak but ubiquitously ex-pressed Gal4 driver tubulin-Gal4 (tub-Gal4) to knockdown gene ex-pression (22). We crossed these Gal4 lines with the individualUAS-RNAi lines to allow expression of shRNAs in different tes-ticular cell types and examined the fertility of the resultingadult male flies. To further characterize the RNAi phenotypes,

1494 | Human Molecular Genetics, 2015, Vol. 24, No. 5


http://www.ensembl.org/





http://flybase.org/

http://flybase.org/

http://flybase.org/

http://flybase.org/

http://flybase.org/

http://www.informatics.jax.org/






http://hmg.oxfordjournals.org/lookup/suppl/doi:10.1093/hmg/ddu557/-/DC1

we also analyzed squashed testes under phase-contrast micros-copy and stained for different markers to distinguish various celltypes within the testes. In total, we screened 29 transgenic RNAilines corresponding to 22 fly genes at two heat-stress tempera-tures (Supplementary Material, Table S2).

Fertility tests revealed seven genes essential for male fertility(Fig. 2A and Table 1): Lar, papi, ERR, CG12404, dmrt11, CG6769 andsfl (Fig. 2C and Supplementary Material, Table S2). Lar is requiredin the germ cells, whereas ERR, CG12404, dmrt11, CG6769 and sflare essential in the somatic cells of the testis. For papi, only indi-viduals expressing tub-Gal4-driven RNAi had reduced fertility(Supplementary Material, Table S2).

Lar is required for spermatogenesis in germ cells

Lar is a transmembrane receptor proteinwith tyrosine phosphat-ase activity that is important for neuron growth and develop-ment (26,27). Interestingly, Lar is also required for oogenesisand spermatogenesis in Drosophila gonads (28,29). Lar is ex-pressed in the GSC membrane at the interface between thehub and the GSCs. Loss of Lar results in loss of GSCs in the testisstem cell niche (29). In this study, we found that Lar is alsoessential for later stages of spermatogenesis in flies. When weused-bam-Gal4 to knockdown Lar expression in 2- to 16-cellspermatogonia, we found that almost all (93.3% at 25°C, N = 60;96.9% at 28°C, N = 96) resulting males were sterile (Fig. 2C).Phase-contrast visualization of squashed adult testes showedthat bam>Lar RNAi testes (N = 26, T = 25°C) had defective elongat-ing and elongated spermatids and no mature sperm (Fig. 3Gand H), indicating that Lar plays a role in the late stage of sperm-atogenesis. In bam>Lar RNAi testes, some cysts close to the testisapex were vasa negative, indicating a loss of germ cells (Fig. 3Aand D). Bam-positive spermatogonial regions were also reduced

in size and closer to the testis tip than those of controls (Fig. 3Band E). In addition, more Eya-positive mature cyst cells accumu-lated in Lar RNAi testes than in the controls (see Fig. 3C, F and I),most likely due to enhanced somatic cell proliferation and differ-entiation in compensation for the lost germ cells.

Five genes are required in testiscular somatic cells tosustain male fertility

Somatic cells are critical for germ cell differentiation, prolifer-ation andmaintenance (30). In our screen, we identified the som-atic cell genes CG12404, dmrt11E, CG6769, sfl and ERR as requiredto maintain male fertility. CG12404 encodes a Yip1 domain-con-taining protein that is involved in vesicular transport from theGolgi complex to the endoplasmic reticulum (31). Two independ-ent RNAi lines, CG12404 RNAi1 (#1463 from VDRC) and CG12404RNAi2 (#41989 from Bloomington), were used to target CG12404in our screen. The c729>CG12404 RNAi1males were totally in-fertile (N = 49 at 25°C and N = 24 at 28°C), and 61.9% (N = 63,T = 25°C) and 72.6% (N = 62, T = 28°C) of c729>CG12404 RNAi2males were infertile (Fig. 2C). Defects were observed from theround spermatid stage in c729>CG12404 RNAi1 testes (N = 24,T = 28°C). The spermatozoa interlaced with each other and losttheir typical rapid movement (Fig. 4A and B). dmrt11E encodes atranscription factor-like protein containing a highly conservedDNA-binding DM domain that plays a critical role in sex deter-mination (32). A total of 43.7% (N = 32, T = 28°C) of c729>dmrt11ERNAi males were infertile, and no mature sperm were observedin the squashed testes (N = 13, T = 28°C). Mild defects at theround spermatid stage were detected (Fig. 4A and C), indicatingthat dmrt11E is required for a late stage of spermatogenesis.CG6769 encodes a zinc finger protein that most likely bindsto RNA. Its physiological function is unknown. 84.0% (N = 50,

Figure 1. A strategy for identifying genes essential for male fertility by GWAS of NOA followed by RNAi-mediated large-scale functional screening in Drosophila. (A) Flowchart of the screening strategy. The selection criteria for candidate genes in Drosophila include a tSNP with an association P-value < 10−5 and a P-value≥ 5 × 10−8 after

multiple validations, Class I or Class II distance between the gene locus and the tSNP (Class I: genes with the tSNP inside the gene body; Class II: adjacent genes

located within the 200 kb flanking the tSNP), homology type (one to many or one to one) and orthology identity >20%. (B) Screening of homologous Drosophila genes

using an in vivo RNAi system. Seven different Gal4 lines were used to drive UAS-RNAi expression under two temperature stress conditions (25 and 28°C). Further

functional analysis followed only if the F1 males were viable and infertile.

Human Molecular Genetics, 2015, Vol. 24, No. 5 | 1495





T = 28°C) of c729>CG6769 RNAimales were sterile. However, thesetestes contained germ cells of all stages, with no obvious mor-phological defects (N = 24, T = 28°C) (Fig. 4A and D).

ERR and sfl are required for testis morphogenesis

ERR is a nuclear receptor that may bind to steroid hormones (33).However, its function is not well understood. In our screen, 56.9%

(N = 58, T = 25°C) and 70.2% (N = 114, T = 28°C) of c729>ERR RNAimales were infertile. We also used c587-Gal4, a cyst cell-specificGal4 driver, to knockdown ERR expression. 45.5% (N = 77, T = 25°C)and 74.6% (N = 59, T = 28°C) of c587>ERR RNAimale flies were ster-ile (Fig. 2C), suggesting that ERR expression in cyst cells is criticalfor male fertility. Interestingly, 88.2% (N = 85, T = 25°C) of thec729>ERR RNAi males and 76.7% (N = 60, T = 25°C) of thec587>ERR RNAi males had abnormal testis morphology

Figure 2. Identification of novel genes required for male fertility in Drosophila. (A) Screen for candidate genes for male fertility in Drosophila. Knockdown of 31.8% of the

candidate genes (N = 22) resulted in infertile males. (B) The number of gene knockdowns that caused lethality at two stress temperature conditions (25°C, black and 28°C,

white). Seven different Gal4 lines were used in the screen. (C) The infertility rates of the Drosophila in which candidate genes were knocked down under two temperature

stress conditions (25°C, black and 28°C, white).

Table 1. Summary of essential fertility genes identified in the screen

Gene (fly) CG No. Chr (fly) Homology type(fly versus human)

Identity (%) Classification Gene (human) Chr (human) Indicating SNPs

papi CG7082 2L One to one 27 Class I TDRKH 1 rs11204885CG12404 CG12404 2L One to many 47 Class II YIPF5 5 rs312620CG6769 CG6769 X One to one 45 Class II ZNF622 5 rs163065sfl CG8339 3L One to many 45 Class II NDST1 5 rs253296Lar CG10443 2L One to many 48 Class I PTPRD 9 rs10978121ERR CG7404 3L One to many 37 Class II ESRRB 14 rs4903393ERR CG7404 3L One to many 36 Class II ESRRG 1 rs2221285ERR CG7404 3L One to many 36 Class II ESRRG 1 rs12023502dmrt11E CG15749 X One to one 26 Class II DMRT2 9 rs10756006

Genes were classified into two classes: Class I, genes with the tSNP located inside the gene body; Class II, adjacent genes located within the 200 kb flanking the tSNPs.



(Fig. 5A1–E1 and Supplementary Material, Fig. S3A), changingfrom the typical long tubular shape to a small and oval-shapedstructure. Upon dissection, all germ cell stages were observednot only in most (75% of c729>ERR RNAi and 92% of c587>ERRRNAi) of the morphologically normal but also in many (18.8% of

c729>ERR RNAi and 85.7% of c587>ERR RNAi) of the morphologic-ally abnormal testes (Supplementary Material, Fig. S1 and S3C).To analyze whether the loss of ERR affects the fate and prolifer-ation of germ cells or somatic cells, we stained the testeswith dif-ferent cellular markers. Germ cells, hub cells and mature cyst

Figure 3. Lar regulates spermatogenesis inDrosophila testes. (A–F) The apex of control (A–C) and bam>Lar RNAi (D–F) testes immunostainedwith anti-Vasa antibody (a germ

cell marker, green in A and D) and anti-FasIII antibody (a hub cell marker, red in A and D). White arrows indicate a cyst of cells without Vasa staining. DE-cadherin marks

hub cells and cyst cells (red in B–C and E–F), and Bam marks spermatogonia (green in B and E). In bam>Lar RNAi testes, the distance between the Bam-positive

spermatogonia and hub cells is smaller than in the control (B and E, white lines). Eya, a marker of mature cyst cells (green in C and F) is increased in bam>Lar RNAi

testes. (G and H) Phase-contrast view of control (G1–G5) and bam>Lar RNAi (H1–H4) testes. Different stages of spermatogenesis in normal testes are indicated,

including spermatogonia (G1), spermatocytes (G2), round spermatids (G3), elongating and elongated spermatids (G4) and mature sperm (G5). The germ cells of

bam>Lar RNAi testes were normal from spermatogonia to the round spermatid stage (H1–H4). Defects in elongating and elongated spermatids were observed, and no

mature sperms were present. (I) Eya-positive cyst cell numbers were increased in bam>Lar RNAi testes. Temperature stress (25°C). Scale bars: 20 µm. (Student’s t-test:

*P-value < 0.05). Error bars represent the SD.





cells were present in the ERR RNAi testes (Fig. 5A2–E2, A3–E3, A6–E6). However, the number of premeiotic germ cells (gonialblasts,spermatogonia and spermatocytes) was strikingly reduced in thec729>ERR RNAi testes with morphological defects (Fig. 5C2 andC4). In addition,more phospho-Histone H3 stainingwas detectedin the ERR RNAi testes than inwild-type controls (Fig. 5A5–E5), in-dicating more cell proliferation in the RNAi testes.

Interestingly, knockdown of sfl in the somatic cells during go-nadal development by esg-Gal4 resulted in oblong-shaped testessimilar to the c729>ERR RNAi or c587>ERR RNAi testes. sfl encodesa glucosamine N-sulfotransferase that is essential for develop-ment and organ morphogenesis (34,35). When we knockeddown sfl by tub-Gal4, esg-Gal4 (T = 28°C) or upd-Gal4, all of theflies died during early stages. However, esg>sfl RNAi male fliesraised at 25°C were viable, but 97.3% (N = 110, T = 25°C) were ster-ile. 51.4% (N = 109, T = 25°C) of the esg>sfl RNAi testes lost thetubular shape and had morphological defects (Fig. 6A1–C1 andSupplementary Material, Fig. S3B). Germ cells at each spermato-genic stage were detected in the squashed testes of mostesg>sfl RNAi males (N = 22 with morphologically defective testesand N = 50 with normal testes) (Supplementary Material, Fig. S2and S3D). Whole-mount testes stained with DNA dye alsoshowed clusters of elongated spermatid nuclei (SupplementaryMaterial, Fig. S4), regardless of whether the testis morphologywas defective. Similar to the c729>ERR RNAi and c587>ERR RNAi

testes, increased phospho-Histone H3 staining was observed(Fig. 6A5–C5). The similarity between the ERR and sfl knockdownanimals suggests that they may work in the same pathway toregulate testis morphogenesis.

DiscussionWith the rapid progress made by GWAS, many risk SNPs havebeen identified in the past few years (5). However, the biologicalsignificance of these SNPs is largely unknown. Considering thehighly heterogeneous human population structure and proper-ties of tSNPs, many SNPs surrounding causal genes may havebeen missed due to low association values and multiple valida-tions. Here, we have presented a systematic screening strategyusing model organisms and human NOA GWAS data.

In this study, of the 22 genes screened, 7 (31.8%) were requiredfor male fertility, indicating the usefulness of functional screen-ing to control for false negatives in the SNP data. It has been es-timated that ∼1500 genes (∼11% of the fly genome) are requiredfor male fertility in Drosophila genome based on several large-scale screen data (14). A 31.8% hit rate is significantly greaterthan what one would expect due to chance (Fisher’s exact testP = 0.008), suggesting that the candidates selected based on ourGWAS data are enriched with genes required for male fertility.In this study, we started with 142 human genes corresponding

Figure 4. CG12404, dmrt11E and CG6769 are required in somatic cells to sustain male fertility. Phase-contrast view of control (A1–A5), c729>CG12404 RNAi (B1–B5),c729>dmrt11E RNAi (C1–C4) and c729>CG6769 RNAi (D1–D5) testes. Different stages of spermatogenesis in control testes are indicated, including spermatogonia (A1),

spermatocytes (A2), round spermatids (A3), elongated spermatids (A4) and mature sperm (A5). Knockdown of CG12404 and dmrt11E by c729-Gal4 caused defects in

round spermatids (black arrows in B3 and C3). No mature sperm were observed in c729>dmrt11E RNAi testes. All stages of germ cells, with no obvious morphological

defects, were observed in c729>CG6769 RNAi testes. Temperature stress (28°C). Scale bars: 20 µm.








to 84 fly orthologs because many human genes did not have flyorthologs. Conservation analysis showed that some humangenes may have multiple orthologs in Drosophila, which makingfunctional testing impossible or difficult. Besides, the sequencesimilarity between some human genes and the Drosophila ortho-logs are very low. Since sequence similarity is an important par-ameter for functional conservation (36,37), only Drosophilaorthologs with sequence identity over 20% were considered inthis study, which is especially important for the orthologous re-lationship of ‘one to many’. Thus, of the 84 Drosophila orthologs,only 38 matched these criteria, and were subjected to the follow-ing functional analysis. The limited homology between thehuman genes and the Drosophila genes indeed make testingevery single candidate impossible. But, comparing with othermodel systems suitable for fast and large-scale analysis such asyeast and worm, Drosophila is the closest to human.

The seven identified fertility genes in flies correspond to eighthuman genes andnine loci associatedwithNOA (Table 1), includ-ing two Class I SNPs (rs10978121 and rs11204885) located insidethe genes PTPRD and TDRKH, the human orthologs of Lar andpapi. In our Drosophila studies, we found that some fertility-

related genes are surrounded by multiple NOA-associated SNPs.For example, fly ERR, a gene encoding an estrogen receptor-related receptor (ESRR) family protein, is essential for malefertility. ERR is an ortholog of human ESRR gamma (ESRRG) andESRR beta (ESRRB), two genes that are surrounded by three NOA-associated SNPs. Multiple NOA-associated SNPs surroundingthe same gene are evidence for its important function in malefertility.

We used seven different Gal4 drivers in our screen, each withdistinct expression timing, expression pattern and strength.Although some have overlapping expression patterns, they didnot produce similar phenotypes in some cases when crossedwith individual RNAi lines. For example, both nos-Gal4 andbam-Gal4 are expressed in germ cells. However, nos>Lar RNAimales were fertile, while the majority of bam>Lar RNAi maleswere sterile. Nos-Gal4 is primarily expressed in the GSC and go-nialblast and is substantially down-regulated at later stages(22), whereas bam-Gal4 expression is restricted to 2- to 16-cellspermatogonia (22,23). In addition to the expression pattern dif-ferences, the expression timing and levels of these Gal4 lines arealso different, whichmight explainwhy the RNAi phenotypes are

Figure 5. ERR is required for testis morphogenesis. (A1–E1) Whole-mount views of testes from control (A1), c729>ERR RNAi (B1 and C1) and c587>ERR RNAi (D1 and E1) flies

by lightmicroscopy. Bothmorphologically normal (B1–B6,D1–D6) and abnormal (C1–C6, E1–E6) testeswere found in c729>ERR RNAi and c587>ERR RNAimales. (A2–E2,A3–E3, A4–E4, A5–E5 and A6–E6) Apical tips of testes immunostained with anti-FasIII (red in A2–E2, A3–E3 and A5–E5), anti-DE-cadherin (red in A6–E6), anti-Vasa (green in

A2–E2), anti-Bam (green in A3–E3) and anti-Eya (green in A6–E6) antibody. Anti-1B1 antibodymarks fusomes (green in A4–E4). The number of premeiotic germ cells (white

arrows in C2 and white double arrows in C4) was remarkably reduced in c729>ERR RNAi testes, which had notable morphological defects. Phospho-Histone H3 (PH3)-

positive cells were increased in number (green in A5–E5) in ERR RNAi testes. Temperature stress (25°C). Scale bars: 100 µmin A1–E1; 20 µm in A2–E2, A3–E3, A4–E4,

A5–E5 and A6–E6.



distinct. In previous studies, Lar was reported to be required forGSC maintenance (29). The lack of fertility defects in the germcells of nos-Gal4 Lar knockdowns are most likely due to lowknockdown efficiency.

Loss of dmrt11E in somatic cells led to spermatogenesis de-fects at late stages. dmrt11E encodes a protein containing a DMdomain that is highly conserved in the Doublesex and mab-3related transcription (DMRT) protein family. Interestingly, DMRTfactor 1 (DMRT1) and DMRT factor 2 (DMRT2), the mammalianorthologs of dmrt11E, map to a region of the genome associatedwith gonadal dysgenesis and XY sex reversal, which makethem candidates formale fertility-required genes (38–40). In add-ition, loss-of-function mutation of DMRT1 has been identified asa risk factor andpotential genetic cause of human spermatogenicfailure (41). Loss of DMRT2 causes neonatal lethality, hinderingits use in the study of fertility in adult animals (42). Our studyprovides direct evidence that DMRT family genes do indeedplay critical roles in male fertility.

In our screen, knockdown sfl or ERR expression in somaticcells resulted in morphological testis defects that phenocopythe drosophila Wnt oncogene analog 2 (DWnt-2) mutant. The forma-tion of a spiral-shaped testis during pupation requires the gonadand genital disc to grow toward and fusewith each other (43). Thegonadal pigment cells and the precursors of smoothmuscle cellsmigrate to form a bilayered sheath that covers the mature gonadand seminal vesicle (43). DWnt-2 is a member of the Wnt family(44). Male flies with DWnt-2 mutations are sterile, with smalland abnormally oblong testes. DWnt-2 is expressed in gonadalsomatic cells and is required formale-specific pigment cell speci-fication and muscle precursor migration (45). Interestingly, sfl isrequired for the biogenesis of heparan sulfate proteoglycans,which are essential for Wg signaling (46). It will be interestingto determine whether it also regulates DWnt-2 signal transduc-tion. ERR encodes a member of the ESRR family, which is part

of the nuclear hormone receptor superfamily. ERR is an orphannuclear receptor whose ligand is not known (47). It has been re-ported that ERR plays a central role in carbohydrate metabolismand directs a critical metabolic transition during development(48). There is no evidence that ESRR family proteins regulateWnt signaling or vice versa. Both ESRR and Wnt family proteinsare involved in oncogenic pathways. It will be interesting to testwhether there is any interplay between ESRR proteins and theWnt signaling pathway.

Interestingly, several of the seven essential fertility genesidentified in this study are highly conserved but have very littlefunctional annotation. For example, there have been few studiesof CG12404 and CG6769 in flies. YIPF5, the human ortholog ofCG12404, is required for endoplasmic reticulum and Golgi struc-ture maintenance in cultured cells (49). ZFN622, the humanortholog of CG6769, has been reported to regulate apoptosis(50). However, the physiological function of both genes isunknown. Here, we found that both genes are required for malefertility in flies. Further investigation will be required to deter-mine the underlying molecular mechanisms through whichCG12404 and CG6769 regulate male fertility.

By exploring the RNA-Seq data set generated by the Geno-type-Tissue Expression (GTEx) project (http://www.gtexportal.org/home/), we found that six (TDRKH, YIPF5, ZNF622, NDST1,PTPRD and ESRRB) out of eight genes identified in our studywere expressed in human testis samples. Although ESRRG andDMRT2 were not present in the GTEx data set, we found thatESRRG is listed as a testis-expressing gene in the UniGene data-base (http://www.ncbi.nlm.nih.gov/unigene/) and DMRT2was re-ported to locate in a region that is critical for testis developmentand is expressed in the adult human testis (40). The presence ofthe genes in human testes suggested that these genes and itsorthologs might be required for male fertility not only in fliesbut also in humans.

Figure 6. sfl is required for testis morphogenesis. (A1–C1) Whole-mount views of testes of control (A1) and esg>sfl RNAi (B1 and C1) flies by light microscopy. Testes with

both normal (B1–B6) and abnormalmorphology (C1–C6) were observed in the esg>sfl RNAimales. (A2–A6, B2–B6 andC2–C6) Apical tips of testes immunostainedwith anti-

FasIII (red in A2–C2,A3–C3 andA5–C5), anti-DE-cadherin (red inA6–C6), anti-Vasa (green in A2–C2), anti-Bam (green inA3–C3), anti-Eya (green in A6–C6), anti-1B1 (green

in A4–C4) and anti-PH3 (green in A5–C5) antibody. Esg>sfl RNAi testes with morphological defects exhibited a reduced number of germ cells (C2).More PH3-positive cells

in number existed in esg>sfl RNAi testes (B5 and C5). Temperature stress (25°C). Scale bars: 100 µm in A1–C1; 20 µm in A2–A6, B2–B6 and C2–C6.



http://www.gtexportal.org/home/






http://www.ncbi.nlm.nih.gov/unigene/







Indeed, it has been reported that some of the genes identifiedin this study have differential expression levels in the NOApatients and the healthy controls (51). The testicular specimensfrom both the NOA patients (n = 18) and the healthy controls(n = 4) express all of the genes identified here. Among them,TDRKH and ESRRB have lower expression levels while YIPF5 hashigher expression level in NOApatients than those in the healthycontrols (P < 0.05) (Supplementary Material, Table S3). Their re-sults were strongly in line with our previous GWAS screeningstudy. So, these expression data in NOA patients showed a possi-bility that these genes are involved in NOA pathogenesis. Thereare high genetic heterogeneities in human. With larger tissuesample size, expression analysis may be able to reveal morehuman fertility-related genes. The elevated YIPF5 level in theNOA patients are surprising because we identified this gene asa fertility-essential gene based on a RNAi knockdown assay. Itwould be interesting to test whether or not YIPF5 overexpressionin flies could lead to male sterility.

The tissue-specific knockdown system we used in this studyis very powerful because it not only bypassed the requirementof the gene in other organs but also allow us to dissect thegene’s function in specific cells and developmental stages ingermline. One concern about this system is that the broadly ex-pressed genes we identified might be essential in other organsthan testis and might not be relevant to NOA, because NOA isasymptomatic in human. Published studies showed pleiotropicnature of many broadly expressed genes in tissues. In flies, theweak alleles of many broadly expressed essential genes leadonly to male sterility (52), suggesting that the broadly expressedgenes could lead to sterility phenotypewithout other symptoms.However, further validating experiments in human can help elu-cidate causal genes of NOA.

In summary, we show that Drosophila RNAi screening is apowerful tool for controlling for false negatives in multiplestage population data. This technique could be further modifiedby adding tissue-specific long-range regulation and expressionquantitative trait loci data (53). The seven essential fertilitygenes we found warrant further study and may help reveal themechanisms underlying NOA.

Materials and MethodsIdentification of Drosophila gene orthologs surroundingtSNPs associated with NOA in humans

The tSNPs associated with human NOA that were considered inthis study all had a P-value < 10−5 in a previous GWAS of NOAper-formed in Han Chinese men (1000 individuals with NOA (cases)and 1703male controls) genotyped for 906 703 SNPswith Affyme-trix Genome-Wide Human SNP Array 6.0 chips (11). To obtaincandidate fertility-related genes, we classified SNP-associatedhuman genes into two classes: Class I, genes that intersectedwith the SNPs and Class II genes located within 200 kb up-stream/downstream of the SNPs. We then selected the corre-sponding unique homologous Drosophila genes (homologytype: orthologous one to many or one to one) of the above asso-ciated human genes for candidate targets to be verified.We onlyconsidered Drosophila genes with at least 20% sequence identitywith their human orthologs. The genome backgrounds forhuman and Drosophila were GRCh37 (hg19) and BDGP R5(dm3), respectively. Gene descriptions and genome localiza-tions for genes and SNPs were batch-obtained from the UCSCgenome browser (http://genome.ucsc.edu/). Orthologousrelationships between human and Drosophila genes were

annotated via BioMart (http://www.biomart.org/) based onEnsembl orthologs.

Statistical methods

Fisher exact tests were conducted in SAS 9.3 (SAS Institute Inc.,Cary, NC) to compare the rate of genes involved in male sterilityin our study and the genome-wide rate estimated by Hackstein’sstudy (14).

The Student’s t test was used for differences between groups.*P-value < 0.05; **P-value < 0.01; ***P-value < 0.001. Error barsrepresent the SD.

Fly stocks and screening crosses

All UAS-RNAi transgenic fly lines were obtained from the VDRC,the Bloomington Drosophila Stock Center (BDSC) and the Tsin-gHua Fly Center (THFC). Gal4 stock lines ordered from BDSCand the Drosophila Genetic Resource Center (DGRC) were as fol-lows: esg-Gal4 (DGRC, #109126), nos-Gal4 (BDSC, #4937), tub-Gal4(BDSC, #5138), c729-Gal4 (BDSC, #6983), upd-Gal4 (BDSC, #26796).Bam-Gal4 and c587-Gal4 were gifts from D.H. Chen and X.Huang, respectively. Canton S flies were used as wild-type flies.We generated two-round crosses and set two temperature stressconditions (25 and 28°C). Transgenic UAS-RNAi males werecrossed to virgin females carrying seven different Gal4 driversat room temperature. To improve the efficiency of RNAi silencing,we employed a higher temperature stress (28°C) after enougheggs had been laid and incubated the eggs at 28°C until adultshatched out. We performed single male fertility tests using a sin-gle F1 RNAi adultmale fly enclosed for 3 days in a crosswith threewild-type virgin females at room temperature.

Dissection and phase-contrast view

Testes were dissected in 1 × phosphate-buffered saline (PBS) andwashed. Shredded testes were observed directly on slides byphase-contrast microscopy after gentle squashing with a coverslip.

Immunofluorescence

Antibodies used were as follows: rat anti-Vasa (DevelopmentalStudies Hybridoma Bank; DSHB, 1 : 20), mouse anti-Eya (DSHB,1 : 20), mouse anti-Fas III (DSHB, 1 : 50), rat anti-DE-cadherin(DSHB, 1 : 20) and rabbit anti-BamC (a gift fromDHChen, 1 : 2000).Secondary antibodies conjugated to A488, Cy3 or A647 (MolecularProbes and Jackson Immunologicals) were diluted at 1 : 1000. Tes-tes were dissected in 1 × PBS and fixed for 30 min in 4% parafor-maldehyde, washed three times in 1 × PBS with 0.1% TritonX-100 (PBST), and blocked for 1 h in 5% bovine serum albumin.The samples were incubated with primary antibodies overnightat 4°C, washed three times for 10 min in 0.1% PBST, incubatedfor 1 h with secondary antibodies at room temperature andthen washed three times for 10 min in 0.1% PBST. The testeswere then stained for 5 min with Hoechst 33 342 (Invitrogen) at1.0 mg/ml. Images were captured on an LSM710 Zeiss confocalmicroscope and processed using Adobe Photoshop CS5 software.

Supplementary MaterialSupplementary Material is available at HMG online.




http://genome.ucsc.edu/





http://www.biomart.org/






Authors’ ContributionsC.T., J.S., X.G. and Z.H. directed the study, obtained financial sup-port and were responsible for the study design, interpretation ofresults andmanuscript writing. C.T., J.S., X.G. and Z.H. performedproject management along with J.Y., H.W., Y.W. and Y.-J.L. Y.W.,T.Z., B.N., Y.L., J.D. and Z.Z. were responsible for data annotationand analysis. J.Y., H.W. and J.-Y.L. were responsible for the Dros-ophila experiments. All authors approved the final version ofthe manuscript.

AcknowledgementsThe authors wish to thank all study participants, research staffand students who assisted with this work. We are grateful toBDSC, THFC, VDRC and DGRC for providing fly lines and DahuaChen (Institute of Zoology, CAS) and XunHuang (Institute of Gen-etics and Developmental Biology, CAS) for sharing reagents andstocks.

Conflict of Interest statement. None declared.

FundingThisworkwas supported by the National Basic Research Programof China (2014CB943100, 2011CB944304, 2012CB966600 and2015CB943003) and the Chinese Natural Science Fund (81222006).

References1. Maduro,M.R. and Lamb, D.J. (2002) Understanding newgenet-

ics of male infertility. J. Urol., 168, 2197–2205.2. Bashamboo, A., Ferraz-de-Souza, B., Lourenço, D., Lin, L., Se-

bire, N.J., Montjean, D., Bignon-Topalovic, J., Mandelbaum, J.,Siffroi, J.P., Christin-Maitre, S. et al. (2010) Human male infer-tility associated with mutations in NR5A1 encoding steroido-genic factor 1. Am. J. Hum. Genet., 87, 505–512.

3. Matzuk, M.M. and Lamb, D.J. (2008) The biology of infertility:research advances and clinical challenges. Nat. Med., 14,1197–1213.

4. Haines, J.L., Hauser, M.A., Schmidt, S., Scott,W.K., Olson, L.M.,Gallins, P., Spencer, K.L., Kwan, S.Y., Noureddine,M., Gilbert, J.R. et al. (2005) Complement factor H variant increases the riskof age-related macular degeneration. Science, 308, 419–421.

5. Johnson, A.D. and O’Donnell, C.J. (2009) An open access data-base of genome-wide association results. BMCMed. Genet., 10,6.

6. Barsh, G.S., Copenhaver, G.P., Gibson, G. and Williams, S.M.(2012) Guidelines for genome-wide association studies. PLoSGenet., 8, e1002812.

7. Panagiotou, O.A. and Ioannidis, J.P. and Genome-Wide Sig-nificance Project. (2012) What should the genome-wide sig-nificance threshold be? Empirical replication of borderlinegenetic associations. Int. J. Epidemiol., 41, 273–286.

8. Galvan, A., Ioannidis, J.P. andDragani, T.A. (2010) Beyond gen-ome-wide association studies: genetic heterogeneity and in-dividual predisposition to cancer. Trends Genet., 26, 132–141.

9. Wang,K., Bucan,M., Grant, S.F., Schellenberg,G. andHakonarson,H. (2010) Strategies for genetic studies of complex diseases.Cell, 142, 351–353.

10. Mathieson, I. and McVean, G. (2012) Differential confoundingof rare and common variants in spatially structured popula-tions. Nat. Genet., 44, 243–246.

11. Hu, Z., Xia, Y., Guo, X., Dai, J., Li, H., Hu, H., Jiang, Y., Lu, F.,Wu,Y., Yang, X. et al. (2012) A genome-wide association study inChinese men identifies three risk loci for non-obstructiveazoospermia. Nat. Genet., 44, 183–186.

12. Chen, H., Liu, Z. and Huang, X. (2010) Drosophila models ofperoxisomal biogenesis disorder: peroxins are required forspermatogenesis and very-long-chain fatty acidmetabolism.Hum. Mol. Genet., 19, 494–505.

13. Hu, Z., Li, Z., Yu, J., Tong, C., Lin, Y., Guo, X., Lu, F., Dong, J., Xia,Y.,Wen, Y. et al. (2014) Association analysis identifies new riskloci for non-obstructive azoospermia in Chinese men. Nat.Commun., 5, 3857.

14. Hackstein, J.H., Hochstenbach, R. and Pearson, P.L. (2000)Towards an understanding of the genetics of human maleinfertility: lessons from flies. Trends Genet., 16, 565–572.

15. Spradling, A., Fuller, M.T., Braun, R.E. and Yoshida, S. (2011)Germline stem cells. Cold Spring Harb. Perspect. Biol., 3,a002642.

16. de Cuevas, M. and Matunis, E.L. (2011) The stem cell niche:lessons from the Drosophila testis. Development, 138, 2861–2869.

17. Rodgers-Melnick, E.B. and Naz, R.K. (2010) Male-biased genesofDrosophilamelanogaster that are conserved inmammaliantestis. Front. Biosci. (Elite Ed), 2, 841–848.

18. Kusz-Zamelczyk, K., Sajek,M., Spik, A., Glazar, R., Jedrzejczak,P., Latos-Biele�nska, A., Kotecki, M., Pawelczyk, L. andJaruzelska, J. (2013) Mutations of NANOS1, a human homo-logue of the Drosophila morphogen, are associated with a lackof germcells in testesor severeoligo-astheno-teratozoospermia.J. Med. Genet., 50, 187–193.

19. Maurano, M.T., Humbert, R., Rynes, E., Thurman, R.E., Hau-gen, E., Wang, H., Reynolds, A.P., Sandstrom, R., Qu, H.,Brody, J. et al. (2012) Systematic localization of common dis-ease-associated variation in regulatory DNA. Science, 337,1190–1195.

20. Ni, J.Q., Zhou, R., Czech, B., Liu, L.P., Holderbaum, L., Yang-Zhou, D., Shim, H.S., Tao, R., Handler, D., Karpowicz, P. et al.(2011) A genome-scale shRNA resource for transgenic RNAiin Drosophila. Nat. Methods, 8, 405–407.

21. Liu, M., Lim, T.M. and Cai, Y. (2010) The Drosophila femalegermline stem cell lineage acts to spatially restrict DPP func-tion within the niche. Sci. Signal., 3, ra57.

22. White-Cooper, H. (2012) Tissue, cell type and stage-specificectopic gene expression and RNAi induction in the Drosophilatestis. Spermatogenesis, 2, 11–22.

23. Singh, S.R., Zhen, W., Zheng, Z., Wang, H., Oh, S.W., Liu, W.,Zbar, B., Schmidt, L.S. and Hou, S.X. (2006) The Drosophilahomolog of the human tumor suppressor gene BHD interactswith the JAK-STAT and Dpp signaling pathways in regulatingmale germline stem cell maintenance. Oncogene, 25, 5933–5941.

24. Gilboa, L. and Lehmann, R. (2006) Soma-germline interac-tions coordinate homeostasis and growth in the Drosophilagonad. Nature, 443, 97–100.

25. Joti, P., Ghosh-Roy, A. and Ray, K. (2011) Dynein light chain 1functions in somatic cyst cells regulate spermatogonial divi-sions in Drosophila. Sci. Rep., 1, 173.

26. Xie, Y., Yeo, T.T., Zhang, C., Yang, T., Tisi, M.A., Massa, S.M.and Longo, F.M. (2001) The leukocyte common antigen-related protein tyrosine phosphatase receptor regulatesregenerative neurite outgrowth in vivo. J. Neurosci., 21, 5130–5138.

27. Chagnon, M.J., Uetani, N. and Tremblay, M.L. (2004) Function-al significance of the LAR receptor protein tyrosine



phosphatase family in development and diseases. Biochem.Cell. Biol., 82, 664–675.

28. Frydman, H.M. and Spradling, A.C. (2001) The receptor-liketyrosine phosphatase lar is required for epithelial planar po-larity and for axis determination within Drosophila ovarianfollicles. Development, 128, 3209–3220.

29. Srinivasan, S., Mahowald, A.P. and Fuller, M.T. (2012) The re-ceptor tyrosine phosphatase Lar regulates adhesion betweenDrosophila male germline stem cells and the niche. Develop-ment, 139, 1381–1390.

30. Kiger, A.A., White-Cooper, H. and Fuller, M.T. (2000) Somaticsupport cells restrict germline stemcell self-renewal and pro-mote differentiation. Nature, 407, 750–754.

31. Kano, F., Yamauchi, S., Yoshida, Y., Watanabe-Takahashi, M.,Nishikawa, K., Nakamura, N. andMurata,M. (2009) Yip1A reg-ulates the COPI-independent retrograde transport from theGolgi complex to the ER. J. Cell. Sci., 122, 2218–2227.

32. Volff, J.N., Zarkower, D., Bardwell, V.J. and Schartl, M. (2003)Evolutionary dynamics of the DM domain gene family inme-tazoans. J. Mol. Evol., 57, S241–S249.

33. Palanker, L., Necakov, A.S., Sampson, H.M., Ni, R., Hu, C.,Thummel, C.S. and Krause, H.M. (2006) Dynamic regulationof Drosophila nuclear receptor activity in vivo. Development,133, 3549–3562.

34. Baeg, G.H., Selva, E.M., Goodman, R.M., Dasgupta, R. and Per-rimon, N. (2004) The Wingless morphogen gradient is estab-lished by the cooperative action of Frizzled and HeparanSulfate Proteoglycan receptors. Dev. Biol., 276, 89–100.

35. Lüders, F., Segawa, H., Stein, D., Selva, E.M., Perrimon, N.,Turco, S.J. andHäcker, U. (2003) Slalom encodes an adenosine3′-phosphate 5′-phosphosulfate transporter essential fordevelopment in Drosophila. EMBO J., 22, 3635–3644.

36. O’Brien, K.P., Remm, M. and Sonnhammer, E.L. (2005) Inpara-noid: a comprehensive database of eukaryotic orthologs.Nucleic Acids Res., 33, D476–D480.

37. Park, C.Y., Wong, A.K., Greene, C.S., Rowland, J., Guan, Y.,Bongo, L.A., Burdine, R.D. and Troyanskaya, O.G. (2013) Func-tional knowledge transfer for high-accuracy prediction ofunder-studied biological processes. PLoS Comput. Biol., 9,e1002957.

38. Raymond, C.S., Kettlewell, J.R., Hirsch, B., Bardwell, V.J. andZarkower, D. (1999) Expression of Dmrt1 in the genital ridgeof mouse and chicken embryos suggests a role in vertebratesexual development. Dev. Biol., 215, 208–220.

39. Raymond, C.S., Shamu, C.E., Shen, M.M., Seifert, K.J., Hirsch,B., Hodgkin, J. and Zarkower, D. (1998) Evidence for evolution-ary conservation of sex-determining genes. Nature, 391, 691–695.

40. Raymond, C.S., Parker, E.D., Kettlewell, J.R., Brown, L.G., Page,D.C., Kusz, K., Jaruzelska, J., Reinberg, Y., Flejter, W.L., Bard-well, V.J. et al. (1999) A region of human chromosome 9p re-quired for testis development contains two genes related toknown sexual regulators. Hum. Mol. Genet., 8, 989–996.

41. Lopes, A.M., Aston, K.I., Thompson, E., Carvalho, F., Gon-çalves, J., Huang, N., Matthiesen, R., Noordam, M.J., Quintela,I., Ramu, A. et al. (2013) Human spermatogenic failure purgesdeleterious mutation load from the autosomes and both sexchromosomes, including the gene DMRT1. PLoS Genet., 9,e1003349.

42. Seo, K.W., Wang, Y., Kokubo, H., Kettlewell, J.R., Zarkower, D.A. and Johnson, R.L. (2006) Targeted disruption of the DM do-main containing transcription factor Dmrt2 reveals an essen-tial role in somite patterning. Dev. Biol., 290, 200–210.

43. Kozopas, K.M., Samos, C.H. and Nusse, R. (1998) DWnt-2, aDrosophila Wnt gene required for the development of themale reproductive tract, specifies a sexually dimorphic cellfate. Genes. Dev., 12, 1155–1165.

44. Russell, J., Gennissen, A. and Nusse, R. (1992) Isolation andexpression of two novel Wnt/wingless gene homologues inDrosophila. Development, 115, 475–485.

45. Baylies, M.K. and Bate, M. (1996) Twist: a myogenic switch inDrosophila. Science, 272, 1481–1484.

46. Lin, X., Buff, E.M., Perrimon, N. and Michelson, A.M. (1999)Heparan sulfate proteoglycans are essential for FGF receptorsignaling duringDrosophila embryonic development. Develop-ment, 126, 3715–3723.

47. Busch, B.B., Stevens, W.C. Jr., Martin, R., Ordentlich, P., Zhou,S., Sapp, D.W., Horlick, R.A. and Mohan, R. (2004) Identifica-tion of a selective inverse agonist for the orphan nuclear re-ceptor estrogen-related receptor alpha. J. Med. Chem., 47,5593–5596.

48. Tennessen, J.M., Baker, K.D., Lam, G., Evans, J. and Thummel,C.S. (2011) The Drosophila estrogen-related receptor directs ametabolic switch that supports developmental growth. CellMetab., 13, 139–148.

49. Yoshida, Y., Suzuki, K., Yamamoto, A., Sakai, N., Bando, M.,Tanimoto, K., Yamaguchi, Y., Sakaguchi, T., Akhter, H., Fujii,G. et al. (2008) YIPF5 andYIF1A recycle between the ER and theGolgi apparatus and are involved in the maintenance of theGolgi structure. Exp. Cell. Res., 314, 3427–3443.

50. Seong, H.A., Jung, H., Manoharan, R. and Ha, H. (2011) Positiveregulation of apoptosis signal-regulating kinase 1 signalingby ZPR9 protein, a zinc finger protein. J. Biol. Chem., 286,31123–31135.

51. Malcher, A., Rozwadowska, N., Stokowy, T., Kolanowski, T.,Jedrzejczak, P., Zietkowiak, W. and Kurpisz, M. (2013) Potentialbiomarkers of nonobstructive azoospermia identified inmicro-array gene expression analysis. Fertil. Steril., 100, 1686–1694.

52. Fuller, M.T. (1993) Spermatogenesis. In: Bate, M., (ed.), TheDevelopment of Drosophila melanogaster. Cold Spring HarborLaboratory Press, Vol. 1, pp. 71–147.

53. Smemo, S., Tena, J.J., Kim, K.H., Gamazon, E.R., Sakabe, N.J.,Gómez-Marín, C., Aneas, I., Credidio, F.L., Sobreira, D.R., Was-serman, N.F. et al. (2014) Obesity-associated variants withinFTO form long-range functional connections with IRX3.Nature, 507, 371–375.



identification of seven genes essential for male fertility through a

Documents