The importance of a halotyrosine dehalogenase for Drosophila fertility

The ability of iodotyrosine deiodinase to salvage iodide from iodotyrosine has long been recognized as critical for iodide homeostasis and proper thyroid function in vertebrates. The significance of its additional ability to dehalogenate bromo- and chlorotyrosine is less apparent, and none of these functions could have been anticipated in invertebrates until recently. Drosophila, as most arthropods, contains a deiodinase homolog encoded by CG6279, now named condet (cdt), with a similar catalytic specificity. However, its physiological role cannot be equivalent because Drosophila lacks a thyroid and its associated hormones, and no requirement for iodide or halotyrosines has been reported for this species. We have now applied CRISPR/Cas9 technology to generate Drosophila strains in which the cdt gene has been either deleted or mutated to identify its biological function. As previously shown in larvae, expression of cdt is primarily limited to the fat body, and we now report that loss of cdt function does not enhance sensitivity of the larvae to the toxic effects of iodotyrosine. In adult flies by contrast, expression is known to occur in testes and is detected at very high levels in this tissue. The importance of cdt is most evident in the decrease in fertility observed when either males or females carry a deletion or mutation of cdt. Therefore, dehalogenation of a halotyrosine appears essential for efficient reproduction in Drosophila and likely contributes to a new pathway for controlling viability in arthropods.

The ability of iodotyrosine deiodinase to salvage iodide from iodotyrosine has long been recognized as critical for iodide homeostasis and proper thyroid function in vertebrates. The significance of its additional ability to dehalogenate bromo-and chlorotyrosine is less apparent, and none of these functions could have been anticipated in invertebrates until recently. Drosophila, as most arthropods, contains a deiodinase homolog encoded by CG6279, now named condet (cdt), with a similar catalytic specificity. However, its physiological role cannot be equivalent because Drosophila lacks a thyroid and its associated hormones, and no requirement for iodide or halotyrosines has been reported for this species. We have now applied CRISPR/ Cas9 technology to generate Drosophila strains in which the cdt gene has been either deleted or mutated to identify its biological function. As previously shown in larvae, expression of cdt is primarily limited to the fat body, and we now report that loss of cdt function does not enhance sensitivity of the larvae to the toxic effects of iodotyrosine. In adult flies by contrast, expression is known to occur in testes and is detected at very high levels in this tissue. The importance of cdt is most evident in the decrease in fertility observed when either males or females carry a deletion or mutation of cdt. Therefore, dehalogenation of a halotyrosine appears essential for efficient reproduction in Drosophila and likely contributes to a new pathway for controlling viability in arthropods.
The flavoprotein iodotyrosine deiodinase (IYD) 2 was discovered first in humans while studying the biochemical basis of thyroid disease (1,2). This enzyme is necessary for maintaining iodide homeostasis and acts by salvaging iodide from iodotyrosines (I-Tyr and I 2 -Tyr, Scheme 1) that are formed during thyroxine (tetraiodothyronine, thyroid hormone) biosynthesis. This process was originally presumed unique to the phylum Chordata because only these organisms are known to generate thyroxine and hence require iodide. In mammals, IYD is primarily expressed in the thyroid gland where this hormone is produced but it is also expressed at lower levels in the digestive tract, kidney, and liver (3) where it may also dehalogenate chloro-and bromotyrosine (Cl-Tyr, Br-Tyr) (Scheme 1) (4,5). These additional halotyrosines are generated during inflammation by activation of myloperoxidase (6,7). Human IYD obtained by heterologous expression demonstrates a nearly equal ability to bind and dehalogenate Br-Tyr and I-Tyr (8). Cl-Tyr also binds with similar affinity but is dehalogenated at a lower rate (20-fold) (8). Whether these additional activities of IYD are necessary or incidental in humans has yet to be determined.
The historical link between IYD and thyroid function might have suggested a possible coevolution. However, genes encoding IYD homologs now appear independent of thyroxine biosynthesis and are common to most animals as well as certain eubacteria and archaea (9,10). This distribution suggests that IYD may play a variety of different roles in biology and the prevalence of its substrates, halotyrosines, may be underestimated (11,12). To date, over 10 homologs of IYD from diverse phyla have been expressed and purified and all of these promote deiodination of I 2 -Tyr (9). Honey bee was selected as the first test of an IYD from arthropods based on its relatively low molecular weight and minimal Cys content. After its deiodinase activity was confirmed (9), attention has turned to Drosophila melanogaster for learning the biological role of IYD in an organism that is not known to require iodide (13,14). Drosophila offers an excellent and well-established model organism but was not initially considered due to complications from alternative splicing of the gene locus (Fig. S1). The IYD gene was originally listed in flybase as CG6279 and has since been named condet (cdt), in honor of Dr. Jean François Condet who discovered that ingestion of iodide could reduce goiter (15). Isoform A of cdt encodes a 484-amino acid N-terminal sequence of unknown structure and function that extends from the canonical IYD domains. This region is common to all Drosophila species sequenced to date but is not shared by other Diptera (16). Isoform B contains only the three domains associated with most metazoan IYDs: an N-terminal membrane anchor, an intermediate region, and a C-terminal active site structure (Fig. S1) (10,17).  (15). Its profile of activity is very similar to that of human IYD. For example, both have high affinities for I-Tyr, Br-Tyr, and Cl-Tyr with K d values ranging from 0.1 to 0.6 M (15, 18). Turnover efficiencies for I-Tyr and Br-Tyr are also nearly equivalent (about 7-9 ϫ 10 3 M Ϫ1 s Ϫ1 ) (8,15). Dehalogenation of Cl-Tyr is less efficient by about 3-and 20-fold for Drosophila and human IYD, respectively. No IYDs have yet demonstrated an ability to dehalogenate fluorotyrosine (10,19) and thus only I-Tyr, Br-Tyr, Cl-Tyr, and their derivatives are possible substrates in vivo. An iodide requirement for Drosophila has yet to be reported but feeding studies with [ 131 I]iodide have established that iodide can accumulate in the protein fraction of cuticle and may also generate I-Tyr and I 2 -Tyr (20). Other halotyrosines have also been identified in the scleroproteins of numerous invertebrates (21,22) and likely result from a peroxidase activity associated with cuticle sclerotization (23). A need to recover bromide or chloride from Br-Tyr and Cl-Tyr would be even more surprising than a corresponding salvage of iodide. Both bromide and chloride are required by Drosophila but are readily available in the environment in contrast to iodide (24 -26).
Existing precedence was not sufficient for speculating on the role(s) of IYD in invertebrates and its expression pattern in Drosophila only adds to its intrigue. Levels of cdt mRNA increase from larvae to adult where it is detected most abundantly in fat body and testis, respectively (27,28). CRISPR/Cas9 (29) has now been used to create Drosophila mutant strains that either have a deletion at this gene locus (cdt ⌬ ) or contain a point mutation E154Q (cdt Q ) with low catalytic activity (15) as an initial effort to learn the biological role of IYD in an organism not known to require iodide. As described below, a lack of the dehalogenase during the larval stage did not sensitize Drosophila to the toxic effects of I-Tyr. Thus, its function in juveniles remains to be established. However, high levels of cdt expression have now been confirmed in testes using in situ hybridization and loss of the protein or its dehalogenase activity reduces fertility. Surprisingly, these effects are observed when either the adult male or female lack sufficient dehalogenase activity.

Generation of Drosophila strains with limited IYD activity in vivo
The biological function of an enzyme can often be defined on the basis of the phenotypes generated by mutants with compromised activity. CRISPR/Cas9 technology has revolutionized such loss-of-activity studies through its ability to remove or alter genes as desired relatively easily in most organisms. To delete the dehalogenase gene cdt (chromosome 3L, bases 11093371 to 11096245) from Dosophila (30), its sequence was analyzed for potential CRISPR sites using the "CRISPR Optimal Target Finder" algorithm (29). From this, a pair of CRISPR target sites that flanked over 90% of the cdt gene was selected (Fig.  S1B). These sites also minimized the chance of potential offtarget reactions that were estimated to be fewer than seven and all contain at least 3 mismatches to the designed gRNAs. A single-stranded oligonucleotide (ODN1, Fig. S2) was created to template formation of 60 bp that flank the Cas9 cut sites for homology-directed repair.
CRISPR/Cas9 was also used to introduce an E154Q mutation at its enzymatic active site to connect a potential phenotype specifically to the dehalogenase activity of IYD. The native Glu-154 participates in substrate recognition and closure of an active site lid that controls catalysis (18,31). Heterologous expression and purification of the E154Q variant had previously been shown to reduce the efficiency of I 2 -Tyr deiodination by more than 800-fold (15). A CRISPR-target site closest to the Glu-154 codon GAG was selected for introducing a single nucleotide mutation to generate a CAG codon for Gln instead (Fig. S1B). This site was only 18 bp upstream of a Cas9 cut site and again allowed for use of a gRNA that had a similarly low probability of off-target reaction (Fig. S1). The oligonucleotide used as the repair template included the desired mutation as well as a silent mutation that destroyed the protospacer adjacent motif of interest to prevent repeated Cas9 cleavage after homology-directed repair (ODN2, Fig. S2).
Single homozygote flies from each of the 35 individual F1 stocks (representing 14 founders with the possible cdt ⌬ genotype) were subject to genomic DNA extraction and PCR screening involving forward and reverse primers that span the dehalogenase domain of cdt (see supporting "Experimental methods" for details). The absence of a 561-bp DNA fragment covering this domain was expected for the deletion mutant. This result was observed in three of the 35 stocks that were examined (Fig. S3A). Additional flanking PCR was performed on these three lines to confirm their genotype. Indeed all three lines generated a 772-bp fragment by PCR in the presence of forward and reverse primers framing the site at which 2580 bp were deleted from cdt (Fig. S3B). The strain illustrated in lane 1 of Fig. S3B was carried forward to examine the phenotypes resulting from gene deletion (cdt ⌬ ).
Three of seven candidates from the F1 progenies generated for the Glu mutation exhibited a diagnostic restriction cleavage with PvuII, indicating a successful introduction of the point mutation (Fig. S4). Subsequent DNA sequencing revealed that two candidates with the desired mutation (Fig. S4, lanes 4 and 5) also contained single nucleotide deletions 11 and 23 bases upstream of the target site, respectively. The third candidate (Fig. S4, lane 6) contained no additional nucleotide insertions, deletions, or mutations from 278 bases upstream to 241 bases downstream of the targeted mutation. This strain was used for studies on the effects of the E154Q mutation (cdt Q ).

A dehalogenase mediates Drosophila fertility A possible metabolic role for IYD in Drosophila
Expression of cdt in the larval fat body suggested a possible protective role for dehalogenation. Feeding studies had previously determined that I-Tyr was toxic to Drosophila larvae (32). This toxicity is apparently derived from the ability of I-Tyr to inhibit tyrosine hydroxylase and ultimately deplete dopamine levels. If tyrosyl residues within proteins became iodinated in Drosophila as they are in vertebrates, then subsequent hydrolysis of these proteins would release I-Tyr. Sensitivity toward I-Tyr might consequently increase if the dehalogenation activity of IYD was compromised. Thus, feeding studies were repeated with the cdt ⌬ and cdt Q strains as well as two controls established by the parent vasa-cas9 and the standard WT (yellow, white) strains.
Addition of 100 M I-Tyr to standard solid Bloomington media had little effect on the ability of Drosophila larvae to develop into adults (Fig. 1). With an increase of I-Tyr to 150 M, a decrease in survival was observed for the strain expressing the active site mutant of IYD (cdt Q ). However, no decrease in viability was noted for the deletion mutant (cdt ⌬ ) or the control flies. Once the I-Tyr supplement was increased to 200 M, survival of larvae to adults decreased in all strains. The cdt Q mutant strain again appeared to be the most sensitive to I-Tyr. The origin of this effect is not yet clear because a complete lack of the catalytic activity (cdt ⌬ ) does not induce a similar sensitivity to I-Tyr. All strains suffered almost complete lethality when the I-Tyr concentration was increased to 300 M. We also confirmed that these results were not a consequence of rejecting the food containing I-Tyr by adding a dye to the media and tracing the food intake. These tests showed that larvae continued to eat the media supplemented with I-Tyr (Fig. S5).
Feeding studies were also repeated with alternative supplements of Br-Tyr and Cl-Tyr. Neither halotyrosine demonstrated the high toxicity of I-Tyr. Significant lethality was not evident for Br-Tyr until its concentrations were increased to 1.5 mM (Fig. 2). Under these conditions, cdt ⌬ inexplicably suffered the least lethality and variability was observed between the two control strains. Again, the presence of Br-Tyr did not discourage consumption of the media (Fig. S5). Only the WT control (yw) strain exhibited a measurable lethality when supplemented with an equivalent concentration of Cl-Tyr. As expected, the products of dehalogenation including Tyr, iodide, and bromide were not detrimental to Drosophila when added to the media at high concentrations (Fig. S6). The insensitivity to excess iodide is consistent with previous studies concluding that iodide did not inhibit a peroxidase-dependent production of melanin (20). Additionally, no adverse effect on survival of Drosophila had previously been observed after exposure to thyroxine or a combination of iodide and tyrosine (33).
In contrast to larvae, adult Drosophila demonstrated no sensitivity to halotyrosines. Survival of adults was unchanged even in the presence of 2 mM I-Tyr (Fig. S7). Equivalent concentrations of Br-Tyr and Cl-Tyr also had no obvious effect. Again, ingestion of a dye included in the food demonstrated that the presence of halotyrosines did not affect the overall eating pattern (Fig. S8). These data along with those from the larvae studies indicate the dehalogenase activity is not essential for detoxifying potential accumulation of halotyrosines that might form during generation or degradation of cuticle. IYD is also not likely required for iodide salvage under standard growth conditions as it is in vertebrates because Drosophila larvae and adults tolerate deletion of IYD well in contrast to humans (34). Attention turned next to the necessity of IYD in fertility because its expression is even higher in testes than in the larval fat body (27,28).

Expression of cdt in testes
Both RNA-seq and Affymetrix-based microarray methods revealed high levels of cdt expression in WT testes compared with levels in ovaries and male carcassed samples (adult males with testes removed) (27, 28). More specifically, WT testes generated 303 reads per kilobase of transcript/million-mapped  reads (RPKM) for cdt followed by always early (aly) mutant testes with RPKM ϭ 107 for cdt (35). By contrast, cdt expression was almost undetectable in bag of marbles (bam) mutant testes (RPKM ϭ 0.12). Because the bam mutation arrests male germ cell differentiation at the mitotic spermatogonial stage (36,37), lack of cdt in bam testes suggests that cdt is not transcribed in mitotic male germ cells (38). By contrast, meiosis is initiated in aly mutant testes but its germ cells are arrested at the G 2 phase as primary spermatocytes (39). The robust expression of cdt in aly mutant and WT testes indicates that it is tightly associated with the meiotic program of the male germ cells.
Antisense RNA probes were designed to detect the N-terminal extension unique to isoform A of cdt and the active site domain common to both isoforms A and B, respectively. Consistent with the RNA-seq results, in situ hybridization in WT testes detected enriched cdt mRNA in spermatocytes but not in the spermatogonial cells (marked with * in Fig. 3, A and B). The region and intensity of staining were also similar for both antisense probes. As controls, the corresponding sense probes produced no detectable signals in WT testes (Fig. 3, C and D). A similar expression pattern in the meiotic male germline has not been reported in humans because no IYD has been detected in gonads or other reproductive organs (3).

Drosophila fertility is promoted by cdt
Fertility assays were first performed for male flies using both cdt ⌬ and cdt Q alleles, due to the high expression of cdt in adult testes. A trans-heterozygous strain carrying the cdt ⌬ allele over a deficiency chromosome (Df) encompassing the cdt locus was also used to avoid any complication due to a second mutation on the same chromosome that may occur if homozygotes are used. A similar trans-heterozygous strain was also generated using cdt Q (cdt Q /Df). Corresponding heterozygotes of cdt ϩ /Df represented a WT control. Each of these fly strains was mated with WT females (vasa-cas9 strain, cdt ϩ /cdt ϩ ) for 24 h and adult progeny were counted on successive days of egg laying (Fig. 4A). Males with one copy of WT IYD (cdt ϩ /Df) yielded an average of 45 adults from eggs laid during the first day. The number of subsequent progeny slowly decreased to an average of 25 adults on the fifth day (Fig. 4B). Males with the deletion of IYD (cdt ⌬ /Df) generated almost the same 38 progeny (average) on the first day as that of the cdt ϩ /Df strain but progeny declined significantly over the following days to an average of only 1 adult progeny on day 5. This represents a 96% loss of fecundity on the last day of observation. In addition, the cdt Q /Df males had an intermediate effect with a 72% loss of fecundity on day 5. Thus, the ability of Drosophila to dehalogenate halotyrosine impacts the fertility of adult males.
Complementary studies mating adult female flies with genotypes of cdt ⌬ /Df, cdt Q /Df, or cdt ϩ /Df with WT males (vasa-cas9 strain, cdt ϩ /cdt ϩ ) under the same protocol generated similar results (Fig. 4C). Again, slight differences in the average progeny were observed on the first day but differences became more  .001 with respect to the control data of (cdt ϩ /cdt ϩ ) x (cdt ϩ /Df). Statistical significance was determined using the Student's t test for a two-tailed distribution of unequal variance.

A dehalogenase mediates Drosophila fertility
obvious when progenies from subsequent days were counted, suggesting an overall decrease in fertility for females with a cdt mutation. This result may be surprising based on the negligible expression of cdt in ovaries (RPKM ϭ 0.25) (35). However, weak expression has been observed in female spermatheca (27), the organ that stores sperm prior to egg fertilization. Thus, cdt expression has the potential to effect sperm in both male and female flies and accordingly, loss of cdt activity in either sex has a significant effect on fertility. The impact is further enhanced by the lack of cdt in both sexes. Mating homozygeous strains either lacking IYD (cdt ⌬ ) or containing its mutant form (cdt Q ) generated even fewer progeny than that observed when only one partner contained a deficiency in cdt (Fig. 5).
In all examples, dehalogenation plays an unanticipated role in Drosophila reproduction. Human IYD is not directly associated with fertility and thus the biological function of Drosophila and human IYD are very different despite the similarities in their specificities of halotyrosine dehalogenation (15,18). Expression of IYD in other insects such as a mosquito (Anopheles gambiae) is not principally observed in testes and is instead distributed widely through most organs at low levels (41). Still, this may be sufficient to affect reproduction in Diptera because even low levels of IYD in female Drosophila are sufficient to maintain fertility when mated with males lacking cdt.
Drosophila IYD has the potential to dehalogenate I-Tyr, Br-Tyr, and Cl-Tyr in vivo but neither its ability to detoxify I-Tyr nor salvage halides likely represents its primary function as suggested by the feeding studies above. Instead, the contribution of IYD to reproduction lends support to an intriguing proposal that halotyrosines may represent a progenitor of thyroxine by functioning as a hormone (42). In both examples, signaling would be dependent on an organism's ability to both halogenate and dehalogenate the parent compounds. Thyroxine controls more than just the metabolic rate in mammals and is essential for proper neonatal development (43). Thyroxine is also crucial for signaling the metamorphosis of amphibians as illustrated most famously by conversion of tadpoles to frogs (44,45). Limited examples have similarly demonstrated that iodinated tyrosines, thyroxine, and thyroxine-like compounds may effect metamorphosis of invertebrates. For example, thyroxine acquired through diet appears to accelerate development of sea hare, sea urchin, and sand dollar (46,47). Most interestingly, I 2 -Tyr appears to be generated endogenously in jellyfish for regulating strobilation of its polyp into free-swimming medusa (48).
Although a role for halotyrosine and cdt in sperm maturation is possible, an additional hypothesis would be necessary to explain the decrease of fertility observed when cdt mutations are only present in female Drosophila (Fig. 4C). A more appealing alternative would have a common basis in both sexes such as maintaining the viability of sperm during development in testes and storage in spermatheca. For example, sperm are highly sensitive to reactive oxygen species and at least in the spermatheca, their longevity appears to depend on controlling metabolism and preventing accumulation of these species (49,50). Such an activity is consistent with the time-dependent loss of fertility observed for cdt mutants and provides a testable hypothesis for future investigations on the details of cdt and halotyrosines in Drosophila. The link between cdt and fertility should also inspire broad investigations into the general requirement of halotyrosines and IYD homologs throughout the phylum Arthropoda.

General
Standard procedures were used for all cloning protocols. ODNs were purchased from IDT (Coralville, IA). Drosophila were raised using standard Bloomington medium at 25°C unless specified. 3-Bromo-L-tyrosine (95% pure) was purchased from AEchem Scientific Corporation (Naperville, IL). 3-Fluoro-L-tyrosine (97% pure) was purchased from Astatech Inc. (Bristol, PA). 3-Iodo-L-tyrosine (97% pure) was purchased from Acros Organics (Waltham, MA). Solutions of 3-halotyrosines and Tyr were prepared in 0.1 N HCl and concentrations were determined using published extinction coefficients (21). Aqueous sodium hydroxide was used to neutralize the media after addition of the acidic halotyrosine solutions.

CRISPR/Cas9-based deletion and mutation of the Drosophila gene cdt
Sense and antisense ODNs encoding the gRNAs (Fig. S1) were cloned separately into the pCFD3-U6:3-gRNA vector using published protocols (http://www.crisprflydesign.org/) 3 (51). The resulting pair of gRNA encoding vectors (250 ng/l each) for targeted gene deletion and a single-stranded ODN template (100 ng/l) were co-injected into Drosophila embryos expressing Cas9 in the germline (BDSC number 51323, vasa-cas9 (29)) by Bestgene Inc. (Chino Hills, CA). To generate the E154Q mutant of cdt, a single gRNA encoding vector (500 ng/l, Fig. S1) and a single-stranded ODN repair template (100 ng/l) that coded for E154Q were similarly co-injected in Drosophila embryos (vasa-Cas9) by Bestgene Inc. The resulting larvae were mated with a MKRS/TM6B balancer and males from the third generation (F3) were screened by PCR (see supporting "Experimental Methods" for details). Strains with the  Fig. 4 were repeated for the designated genotypes. The box plots are defined by the 25th and 75th percentiles. The midline indicates the average and the bar indicates standard deviation. ***, indicates a p value Ͻ0.001 with respect to the control data of (cdt ϩ /cdt ϩ ) ϫ (cdt ϩ /cdt ϩ ). Statistical significance was determined using the Student's t test for a two-tailed distribution of unequal variance.
A dehalogenase mediates Drosophila fertility desired homozygous genotypes (cdt ⌬ or cdt Q ) were confirmed by DNA sequencing. A control strain containing the WT gene (cdt ϩ ) was generated with equivalent crosses between adults from the MKRS/TM6B balancer strain and those from the parent vasa-cas9 strain. The Drosophila strain yellow, white (yw), commonly used as a reference in laboratory experiments, was also employed as a control for all feeding experiments (53).

Larval feeding
Mature egg laying adults of the desired genotype (homogzyous) were transferred to vials containing apple juice/agar medium (40% Giant brand apple juice, 1.5% agar, and 0.05% Tegosept) and allowed to lay eggs. Adults were removed after 24 h and L1 larvae were collected between 24 and 30 h after egg laying. Fifty L1 larvae were transferred to each vial containing the standard feeding media, FD&C blue 1 dye (0.007% w/v), and the indicated supplement. Food consumption by the larvae was detected through ingestion of the dye as described above and using larvae raised solely on the standard Bloomington media containing halotyrosine and dye (Fig. S5). Vials were maintained at 25°C and the number of larvae surviving to adulthood was quantified.

Adult feeding
One-day-old male and female adult flies (25 each) were transferred to a vial containing a tissue (KimWipe, Kimberly Clark) soaked in 3 ml of an apple juice/halotyrosine feeding mixture (75% (v/v) apple juice, 2 mM halotyrosine, 0.05% (w/v) Tegosept, and 0.05% (w/v) FD&C blue 1 dye). Consumption of the liquid mixture by adult flies was confirmed through ingestion of the dye (Fig. S8). Adult Drosophila were maintained on apple juice/halotyrosine mixtures at 25°C for 5 days and the number of surviving adults was quantified.

In situ hybridization to identify cdt expression
Expression of WT cdt in testes (yw strain, whole mount) was detected by RNA in situ hybridization using standard procedures (40,54). Briefly, primers 5Ј-CATGAAGTTTCGGTAG-AAGAG-3Ј (forward) and 5Ј-CATCTGCACTTGCTGGCTA-TT-3Ј (reverse) were used to amplify a 405-bp region of cdt isoform A from a Drosophila testes cDNA library (35) using Taq DNA polymerase (Thermo Scientific). Similarly, primers 5Ј-ATTGTGGAACAGGAGGAGCTG-3Ј (forward) and 5Ј-ATTCTTTCTCGCCAAGTCGGG-3Ј (reverse) were used to amplify a 393-bp region of cdt that is common to both isoforms A and B. The PCR products with 5Ј-A overhangs were then individually ligated into the linearized pGEM-T Easy plasmid (Promega) with 3Ј-T overhangs using T4 DNA ligase (Promega, Madison, WI). Plasmids were purified from single colonies and the desired inserts were subsequently subcloned into the pBluescript II vector using SacII and PstI restriction sites. Transformants containing inserts were identified via blue-white colony screening and the orientation of the inserts was determined by DNA sequencing (Genewiz, South Plainfield, NJ). Based on the orientation of the DNA inserts, antisense and sense hybridization probes were synthesized using T3 and T7 RNA polymerases (Roche Life Science) after linearizing the vector with PstI or SacII, respectively. A digoxigenin (DIG) label was incorpo-rated in the probes through DIG-labeled UTPs (DIG RNA labeling mix, Roche Life Science). All subsequent procedures including hydrolysis of RNA probes, testes preparation, and in situ hybridization were performed as described previously (40,54). Testes were imaged with a Zeiss HXP120C Apotome microscope.

Fecundity after a limited mating period
The balanced deficiency (Df) fly stock for cdt carrying a deletion on chromosome 3L (11070526 to 11247145) was obtained from the Bloomington Drosophila stock center (BDSC number 27578). Homozygous virgin females of genotype cdt ⌬ and cdt Q derived from the vasa-cas9 strain and a WT control (cdt ϩ ) were individually mated with Df/TM6B males. The first generation without the humeral phenotype provided the cdt ⌬ /Df, cdt Q /Df, and cdt ϩ /Df strains. Newly eclosed virgin males and females of each strain were maintained separately for 2-3 days and 3-4 days, respectively, prior to mating. Concurrently, newly eclosed virgin males and females of the cdt ϩ /cdt ϩ genotype were maintained equivalently and then combined individually with the specified strains for 24 h at 25°C. Males were then removed, and the females were transferred to new vials every 24 h for 5 consecutive days to lay eggs. The progeny emerging from each vial was quantified from day 10 to 18 as described previously (52). Vials were excluded from analysis if any of the adults died during this period.