ADAD2 functions in spermiogenesis and piRNA biogenesis in mice

Adenosine deaminase domain containing 2 (ADAD2) is a testis‐specific protein composed of a double‐stranded RNA binding domain and a non‐catalytic adenosine deaminase domain. A recent study showed that ADAD2 is indispensable for the male reproduction in mice. However, the detailed functions of ADAD2 remain elusive.


INTRODUCTION
Spermatogenesis is a complicated, highly coordinated developmental process fundamental for accurate propagation of genetic information to the next generation. 1,2 Transposable elements (TEs), or transposons, are discrete segments of DNA capable of changing their genomic locations, replicating themselves, and integrating into and creating mutations in the genome. 3 TE-mediated mutations are of evolutionary significance, but uncontrolled proliferation of TEs has deleterious effects on host fitness. To attenuate the potential threats of TEs, the male germline recruits a sophisticated, small RNA-based defense system involving P-element-induced wimpy testis (PIWI)-like proteins and PIWI-interacting RNAs (piRNAs) to prevent the propagation of TEs across the genome. 3,4 piRNAs, which are generally 25-31 nucleotides (nt) in length, are generated via the primary and secondary processing pathways. 5 Primary piRNA biogenesis involves the production of precursor piRNAs (pre-piRNAs) by endonuclease-mediated fragmentation of long, singlestranded RNAs that are transcribed from piRNA cluster loci containing retrotransposon sequences. 6,7 The 3′ termini of the pre-piRNAs are subsequently trimmed and 2′-O-methylated. 8 The resultant mature primary piRNAs are characterized by a preference for a uridine (1U) at the first position. Primary piRNAs guide the recognition and PIWImediated cleavage of complementary transcripts between the 10th and 11th nucleotides to generate secondary piRNAs, an amplification process known as the ping-pong cycle. 9 Primary and secondary piR-NAs harbor a 10-nt complementary overlap, and the secondary piRNAs exhibit a bias for an adenine at the 10th position (10A) from their 5′ termini. [10][11][12] Ping-pong-derived piRNAs are abundant in fetal germ cells but are low in adult testes. 4 The mouse genome encodes three testis-enriched PIWI-like proteins, MIWI (Piwi-like protein 1 or PIWIL1), MILI (Piwi-like protein 2 or PIWIL2), and MIWI2 (Piwi-like protein 4 or PIWIL4), that exhibit distinct expression timing and molecular functions during spermatogenesis. [13][14][15] Mili and Miwi2 are initially expressed in the prenatal testes; mutations in MILI or MIWI2 cause spermatogenic arrest at the zygotene/pachytene spermatocyte stage. 14,16 In con-trast, expression of Miwi is restricted to late spermatogenesis, from pachytene spermatocytes to elongating spermatids; ablation of MIWI results in spermiogenic arrest at the round spermatid stage. 3,17 Based on the timing of expression, these PIWI-bound RNAs are referred to as fetal and postnatal pre-pachytene or pachytene piRNAs in spermatogenic cells. 18,19 Pre-pachytene piRNAs in fetal testes are primarily derived from TEs and are associated with MILI and MIWI2. 5,20 Different from pre-pachytene piRNAs, pachytene piRNAs are predominantly transcribed from intergenic regions termed pachytene piRNA clusters, while 10%-20% of them are produced from non-cluster regions such as coding RNAs, non-coding RNAs, repeats, and introns. 18,21 Both cluster and non-cluster-derived pachytene piRNAs bind to MILI and MIWI. 3 Because of their repeat-devoid origin, the detailed functions and target transcripts of pachytene piRNAs remain to be clarified. After meiosis, sperm chromatin undergoes extensive chromatin remodeling, during which histone replacement occurs. 10 The incorporation of histone variants is associated with open and accessible chromatin, resulting in loosened transcriptional control. 10,22 Pachytene piRNAs have been reported to regulate post-transcriptional silencing of mRNAs and lncR-NAs, which are transcribed because of this genome-wide derepression of transcription. [23][24][25][26] In addition to PIWI-like proteins, a multitude of proteins have been discovered to be essential for piRNA biogenesis and function, such as GASZ, 27 31 Similarly, mice lacking ADAD2 or RNF17 show male-specific sterility because of abnormal spermatid differentiation. 34,40 RNF17 is localized to granules distinct from other known nuages in late pachytene and diplotene spermatocytes. 34 Depletion of RNF17 unleashes the ping-pong cycle, which aberrantly produces secondary piRNAs that degrade not only TEs but also mRNAs and lncRNAs. 7 Similar to RNF17, ADAD2 also forms prominent granules in pachytene spermatocytes. 40,41 Nevertheless, it is not clear how ADAD2 governs the differentiation of spermatids. In this study, employing CRISPR/Cas9-based gene editing, transcriptomic, and proteomic techniques, we have investigated the molecular functions of ADAD2 and its potential involvement in piRNA biogenesis.

Animals
Wildtype mice were purchased from Japan SLC, Inc.

Generation of knockout mice
Adad2 null (Adad2 +/-) and mutant (Adad2 +/∆ ) mouse lines and the Adad1 knockout mouse line were generated by the CRISPR/Cas9 system as previously described. 42 The sequences of sgRNAs and mutant alleles are enumerated in Table S1. The sequences of primers for genomic PCR are listed in Table S2.

Reverse transcription polymerase chain reaction
Mouse cDNA was prepared from various tissues of adult mice, or from 5-to 42-day-old mouse testes. Alternatively, cDNA was prepared from spermatocytes and spermatids, which were isolated and purified from mouse testes as previously described. 43 Briefly, testes from 24-day-old mice were dissected to remove the tunica albuginea. The decapsu-  Table S2.

Fertility tests
Sexually mature Adad1 or Adad2 knockout males were individually caged with three B6D2F1 female mice for 8 weeks. During this period, vaginal plugs were examined as an indicator of successful copulation, and the number of offspring in each litter was recorded at birth. Three knockout males were analyzed to meet the requirements for statistical validity. The fecundity of three wildtype B6D2F1 males was tested in parallel as positive controls. After 8 weeks of breeding, the male mice were withdrawn from the cages, and the females were kept for another 3 weeks to allow the final litters to be delivered.
The testis sections were dried on adhesive microscope slides, permeabilized, and blocked with 0.1% Triton X-100, 3% BSA, and 10% goat serum in PBS for 1 h at room temperature. The samples were then incubated with primary antibodies for 3 h at room temperature or overnight at 4 • C. After three washes in 0.1% Triton X-100 in PBS, the sections were incubated with fluorophore-conjugated secondary antibodies for 1 h at room temperature and then stained with 1 μg/mL Hoechst 33342 (ThermoFisher) for 30 min at room temperature. The sections were mounted with Epredia ™ Immu-Mount (Fisher Scientific, Pittsburgh, PA, USA) prior to imaging. Fluorescence images were captured with a Nikon Eclipse Ti microscope equipped with a Nikon C2 confocal module (Nikon, Tokyo, Japan).

Mass spectrometry
Proteomic analysis was performed as previously described. 45

Statistical analyses
Data are presented as mean values and error bars indicate standard deviation (SD). Experimental groups were analyzed statistically using an unpaired two-tailed Student's t-test. p-Values less than 0.05 were considered statistically significant (*p < 0.05; **p < 0.01; ***p < 0.001).

ADAD2 is essential for male fertility in mice
As depicted by reverse transcription polymerase chain reaction (RT-PCR), mouse Adad1 and Adad2 are testis-specific genes initially expressed at postnatal day 5 ( Figure 1A,B). In humans, ADAD2 is also expressed in brain ( Figure S1A). According to a previous single-cell RNA-seq (scRNA-seq) analysis of mouse spermatogenic cells, 48 Adad1 is highly expressed in late spermatocytes and early spermatids, whereas Adad2 shows peak expression in pachytene spermatocytes homozygous null and mutant testis lysate ( Figures 1E and S2C). When consecutively paired with wildtype females, the homozygous mutant males with either a complete deletion or a frameshift allele failed to sire any offspring ( Figures 1F and S2D), demonstrating that Adad2 is essential for male reproduction in mice.
Histological analyses of testis and epididymis sections revealed that spermatogenesis was generally normal in Adad2 -/and Adad2 ∆/∆ males until the early round spermatid stage (Figures 2A and S3A).
Binucleated round spermatids were occasionally observed in the  Figure S3A).
Immunohistochemistry revealed that ADAD2 was detected as granules exclusively localized to the pachytene spermatocytes ( Figure 2B).

ADAD2 interacts with multiple RNA-binding proteins in male germ cells
To acquire a better understanding of the molecular functions of ADAD2, we analyzed its interacting proteins in wildtype testes (with Adad2 -/serving as a negative control) by co-IP/MS ( Figure 4A). GO analyses unveiled that a significant portion of the interacting proteins exhibit RNA-binding ability ( Figure S5A). These include DDX25, MAEL, MILI, MIWI, RNF17, and YTHDC2 ( Figures 4B and S5B), which function in piRNA biogenesis during spermatogenesis. 7,20,31,32,64,65 Noticeably, DDX25 and RNF17, which were detected in the ADAD2 interactome, have been concurrently identified as MAEL-interacting proteins 31 ( Figure S5C). By co-IP tandem Western blot analyses, we confirmed that ADAD2 interacts with ADAD1, MAEL, MILI, MIWI, RNF17, and YTHDC2 ( Figure 4C-E). Immunohistochemistry revealed that ADAD2 co-localizes with RNF17 in large granules in the wildtype pachytene spermatocytes. In Adad2 -/spermatocytes, the numbers of RNF17 granules are markedly decreased ( Figure 4F). The ADAD2 granules show minor co-localization with MILI in spermatocytes ( Figure S6).

Deletion of Adad2 alters piRNA populations in the male germline
Since ADAD2 is associated with multiple RNA-binding proteins implicated in piRNA biogenesis, we carried out deep sequencing analyses of small RNAs in wildtype and Adad2 ∆/∆ postnatal day 24 testes. Length distribution analysis of small RNAs indicated that the abundance of piRNAs ranging from 24 to 31 nt was comparable in wildtype and mutant testes ( Figure 5A). No significant difference was observed in the nucleotide distribution of small RNAs or the ratio of 1U-biased primary piRNAs between wildtype and mutant ( Figure 5A,B). Interestingly, the percentage of secondary piRNAs exhibiting 10th A bias was increased in Adad2 ∆/∆ testes ( Figure 5B). In corroboration with this observation, the mutant showed higher expression of piRNAs carrying 10-nt complementary overlaps, another signature of piRNAs derived from the ping-pong cycle (Adad2 +/+ Z-score = 19.1; Adad2 ∆/∆ Z-score = 33.1; Figure S8A). To clarify the changes in the piRNA populations, we next analyzed the expression of piRNAs transcribed from different genomic loci.
While the expression of intron-, SINE-, LINE-, and LTR-derived piRNAs was significantly increased, cluster-derived pachytene piR-NAs showed decreased expression in the mutant males ( Figure 5C).

DISCUSSION
In this study, we demonstrate the essential role of ADAD2 in male  Figure S3A,B).
Thus, as also pointed out by Snyder et al., 40 such anomalies may lead to a reduced number of round spermatids in Adad2-deficient males, thereby diminishing the relevance of the RNA-seq outcomes. To overcome this potential concern, Snyder et al. 40 focused only on the downregulated spermatocyte-enriched genes and upregulated round spermatid-enriched genes in the mutant testes. This approach, however, neglects a number of downregulated spermatid-enriched genes with significant fold changes (e.g., Prm1, Prm2, Tnp1, and Tnp2). Instead, RT-PCR or quantitative PCR could be alternatively or additionally performed to verify the mRNA levels of genes of interest in purified spermatocytes or spermatids (e.g., Figure 3B-D). In future studies, RNA-seq analyses could be further conducted using testes at 14 days postpartum to elucidate how the spermatocyte-localized ADAD2 underpins subsequent spermatid differentiation.
ADAD2 interacts and co-localizes with RNF17 in spermatocytes ( Figure 4C,F). The partial co-localization of ADAD2 and MILI suggests that the ADAD2/RNF17 granules might be intermitochondrial cement ( Figure S6). Depletion of Adad2 impairs formation of RNF17 granules ( Figure 4F). Consistent with the tight association between ADAD2 and RNF17, mutant mice lacking either protein exhibit arrested spermiogenesis and massive downregulation of protein-coding transcripts ( Figures 3A and S4). 7,34 In both mutants, the expression of total pachytene cluster piRNAs, and MIWI-and MILI-bound pachytene cluster piRNAs, is decreased, whereas the abundance of transposonderived secondary piRNAs is increased ( Figures 5C,D and S8A). 7 In response to such alterations in the piRNA population, mRNAs encoding