ARRDC5 expression is conserved in mammalian testes and required for normal sperm morphogenesis

In sexual reproduction, sperm contribute half the genomic material required for creation of offspring yet core molecular mechanisms essential for their formation are undefined. Here, the α-arrestin molecule arrestin-domain containing 5 (ARRDC5) is identified as an essential regulator of mammalian spermatogenesis. Multispecies testicular tissue transcriptome profiling indicates that expression of Arrdc5 is testis enriched, if not specific, in mice, pigs, cattle, and humans. Knockout of Arrdc5 in mice leads to male specific sterility due to production of low numbers of sperm that are immotile and malformed. Spermiogenesis, the final phase of spermatogenesis when round spermatids transform to spermatozoa, is defective in testes of Arrdc5 deficient mice. Also, epididymal sperm in Arrdc5 knockouts are unable to capacitate and fertilize oocytes. These findings establish ARRDC5 as an essential regulator of mammalian spermatogenesis. Considering the role of arrestin molecules as modulators of cellular signaling and ubiquitination, ARRDC5 is a potential male contraceptive target.

was generated.This is stated clearly in the respective supplementary figure legend, and should be made equally clear in the results section.
Reviewer #2 (Remarks to the Author): The authors have undertaken a comprehensive study where they compare the testis transcriptomes across three species.This will provide useful information for use by the field.They have then selected one highly conserved gene, ARRDC5, to test the function of.ARRDC5 is required for male fertility and appears to play a role specifically in elongating sperm development.The phenotype resulting from the loss of ARRCD5 is reasonably well described (noting comments below).The data is broadly of a high quality and the manuscript well written.I do however note, and as detailed below, there are several places where the person/people who wrote the manuscript and prepared the figures have made small but significant errors.The figures do not always reflect what is written in the text.
Of greater significance, I wonder why germ cell ubiquitination was not studied to formally test the role of ARRDC5 in ubiquitination?Doing so would add some mechanistic data to the manuscript.Specific comments Title -I suspect that the term 'evolutionarily conserved' will bother career evolutionary biologists.As such I suggest that the title be changed to something like 'Arrdc5 is a conserved gene across mammals and is required for male fertility'.
Line 21 -OAT is technically not a diagnosis but rather a description of a clinical presentation.
Line 28 -add in the word mammalian before cell as there are other examples in other species.
Line 53 -the energy required for sperm tail function is provided by mitochondria (as stated) but also by the glycolysis pathway anchored to the fibrous sheath.The latter is absolutely critical in mouse sperm.Please update this sentence In several places -use of the word stage should be checked, noting the strict meaning of 'stage' in spermatogenesis.
Fig. 4 -a tip.Staining sperm with haematoxylin and eosin is cheap and gives very nice results.I personally prefer it to Dip quick in that you can see more sub-structures.
Line 234 -delete the word syndrome as it is used incorrectly.Line 237 -this section needs to be rewritten although I am not questioning the result.Capacitation and hyperactivation are biochemically separate processes.They usually occur in parallel, but they can be uncoupled.The ability to undergo the acrosome reaction is directly coupled to capacitation.Line 251 -reword.PNA is not an acrosome component, rather it binds to components of the acrosome.On a related topic, did you check that KO sperm have acrosomes in the first place ie.did nonionophore challenged sperm stain positively with PNA?The IVF data could be consistent with the absence of an acrosome.Given that some of the IF data suggests ARRDC5 is localised to the acrosome region, this should be checked.Ditto while I can see some hooked nuclei in panel A -/many sperm appear to have blunted nuclei suggesting an acrosome defect.Some high magnification sperm images (stained with H&E) might be helpful.The blunted heads can also be seen in 6B.Unfortunately, the TEM in 6C is not informative in relation to this question.Fig. 5C -I found this an odd way to express the data.How was stage defined?How can you have reduced stage II but normal stage III.Personally, I would remove this panel and concentrate on describing/illustrating this pathology.Can I also suggest that someone with a lot of experience check the staging indicated in 5A -I disagree with some of the labels.Fig. 6C -please indicate the defects listed in the text in the figures e.g. it is not possible to see misplaced axonemes, disorganised mitochondrial sheaths or the absence of the post-acrosomal region in the images chosen.These are key claims.
Line 326 -the images in Fig. 7C do not support that statement that there are multiple nuclei and a single axoneme.Remember that TEM shows a thin 2D section of a 3D object.Conclusion -I disagree with the statement that few testes specific genes when knockout have been shown to cause male sterility in isolation.There would be several hundred.Perhaps rewrite the sentence to more accurately reflect previous data.I appreciate this is a promising phenotype.
Reviewer #3 (Remarks to the Author): In this manuscript, Giassetti et al., report an essential role of arrestin-domain containing 5 (Arrdc5) in mammalian spermatogenesis.The authors conducted single cell RNA-seq analyses of testicular cells from prepubertal mice, pig, and cattle and found that Arrdc5 is only expressed in germ cells in these species.Integrated analyses using ENCODE and human GTEx dataset suggest that this gene is testis specific.The authors then generated Arrdc5-eGfp and Arddc5-/-mouse lines to provide evidences that Arrdc5 is present in spermatids and plays an indispensable role in spermiogenesis.Development of spermatogonia, spermatocytes and haploid round spermatids are not disrupted, however, sperm morphology and function are severely affected by Arrdc5 deletion.The author concluded that Arrdc5 is a core regulator of spermatogenesis and a potential male contraceptive target.Although this study provides important data regarding the role of Arrdc5 in spermiogenesis, it contains significant flaws that fail to support major conclusion of this manuscript.
1.Although not published in peer reviewed journal, a publicly available report (International Mouse Phenotyping Consortium, MGI: 1924170) indicates that, in addition to male infertility, Arrdc5 knockout mice show abnormal retina morphology, increased hemoglobin, circulating fructosamine and other metabolites.The authors stated that loss of Arrdc5 causes male sterility without impacting other physiological processes, but did not provide any experimental evidence.For example, it is important to examine eye development because a male contraceptive target must not affect other important organs.2. "testis specific" and "testis enriched" are completely different.Human GTEx data show that Arrdc5 is enriched in testis, but also present in whole blood and other tissues (Supplemental Fig. 4).Unfortunately, whether these tissues (especially whole blood and retina) of mice, pig and cattle express Arrdc5 is not clear.Therefore, the conclusion that Arrdc5 in mammalian species from mice to human is germ cell specific lacks important evidence.3. A big shortcoming is that this study only describes the infertile phenotype, although the authors state that Arrdc5 serves as a "core regulator".Without defined molecular mechanism, it is difficult to interpret the current data as "core" regulator of spermatogenesis.Similar defects have been reported in other knockout lines, is Arrdc5 functionally interacting with these proteins such as Capza3, or Ccdc62 to serve its core role? 4. SSC transplantation experiment was not necessary for this study.Only one sperm is shown in Fig. 8B and it appears that elongating spermatids were entirely missing and SSC transplantation generated a completely different phenotype.
Author Response: We agree with the reviewer that presence of multiple axonemes cannot be observed from the 2D section produced by TEM.Thus, the sentence has been revised to remove mention of this possible oddity.
Author Response: We thank the reviewer for this comment and agree that the data are important for demonstrating that impaired sperm morphogenesis of Arrdc5-/-mice is intrinsic to the germ cells.
Reviewer Comment 16: Conclusion -I disagree with the statement that few testes specific genes when knockout have been shown to cause male sterility in isolation.There would be several hundred.Perhaps rewrite the sentence to more accurately reflect previous data.I appreciate this is a promising phenotype.
Author Response: We appreciate the reviewer's point but are unclear as to which statement in the discussion implies that few testes specific genes cause male sterility when knocked out.The one sentence that the reviewer may be referring to is, "Few studies have uncovered genes expressed specifically in testicular germ cells that when inactivated lead to male specific sterility and even less have associated the expression of essential regulators identified in mice to evolutionary conservation in other mammalian species".This statement is made to indicate that knowledge of genes expressed specifically in testicular germ cells with a critical role in spermatogenesis is limited.We certainly agree with the reviewer that a multitude of genes with testis-specific expression are known but believe that assignment to germ cells as intrinsic regulators of spermatogenesis is largely undefined.As suggested by the reviewer, we have attempted to rewrite the statement to more accurately reflect the state of understanding for the field of spermatogenesis.

Response to Reviewer 3
Summary Comments by the reviewer: In this manuscript, Giassetti et al., report an essential role of arrestin-domain containing 5 (Arrdc5) in mammalian spermatogenesis.The authors conducted single cell RNA-seq analyses of testicular cells from prepubertal mice, pig, and cattle and found that Arrdc5 is only expressed in germ cells in these species.Integrated analyses using ENCODE and human GTEx dataset suggest that this gene is testis specific.The authors then generated Arrdc5-eGfp and Arddc5-/mouse lines to provide evidences that Arrdc5 is present in spermatids and plays an indispensable role in spermiogenesis.Development of spermatogonia, spermatocytes and haploid round spermatids are not disrupted, however, sperm morphology and function are severely affected by Arrdc5 deletion.The author concluded that Arrdc5 is a core regulator of spermatogenesis and a potential male contraceptive target.Although this study provides important data regarding the role of Arrdc5 in spermiogenesis, it contains significant flaws that fail to support major conclusion of this manuscript.
Reviewer Comment 1: Although not published in peer reviewed journal, a publicly available report (International Mouse Phenotyping Consortium, MGI: 1924170) indicates that, in addition to male infertility, Arrdc5 knockout mice show abnormal retina morphology, increased hemoglobin, circulating fructosamine and other metabolites.The authors stated that loss of Arrdc5 causes male sterility without impacting other physiological processes, but did not provide any experimental evidence.For example, it is important to examine eye development because a male contraceptive target must not affect other important organs.
Author Response: We appreciate the reviewer's comment and became aware of the non peerreviewed publicly available report of Arrdc5 knockout mice after we had generated the CRISPR edited knockout described in the current manuscript.Although the standard IMPC phenotyping pipeline reported that the Arrdc5 knockout mice generated by the Wellcome Trust Sanger Institute (WTSI) using ES cell targeting have an abnormal retina morphology, we have not observed overt defects in vision for any animals of the knockout line generated with CRISPR-Cas9 gene editing.Although our knockout animals have seemingly normal health, except for male specific sterility, which suggests that lack of ARRDC5 does not negatively impact other organs, we have not carried out in-depth assessment of all physiological systems.Thus, we have tempered the conclusionary statement to indicate that the findings demonstrate that ARRDC5 function is required specifically for male fertility and suggest that deficiency does not have overt negative impact on other physiological systems that impact health.
Regarding the reviewer's suggestion to examine eye development, we disagree that this assessment is needed for the current study for the following reasons: 1. New RT-PCR data has been included in the revised manuscript that demonstrates Arrdc5 gene expression is absent in the eye of adult mice (Figure 2A in the revised manuscript, formerly Figure 2C).Although this finding does not assess whether Arrdc5 is expressed at any point in fetal or perinatal development of the eye, it does demonstrate that expression is undetectable in the adult phase of life which is when a male contraceptive would be applied.2. As best we can tell, mice from our knockout line do not have obvious visual impairment compared to wild-type littermates.Even out to over 1 year of age, the knockout animals retain seemingly normal eyesight and gross abnormalities of the eye are not observable.3. We have included new body weight data for Arrdc5 knockout mice at 3 and 16 weeks of age that shows no measurable difference compared to wild-type littermates (Supplementary Figure 10E).These findings are a general indicator of normal physiology for Arrdc5 knockout mice.If the vision and/or metabolism were negatively affected by lack of ARRDC5 expression, one would expect a difference in body weight by 4 months of age compared to wild-type counterparts.
Reviewer Comment 2: "testis specific" and "testis enriched" are completely different.Human GTEx data show that Arrdc5 is enriched in testis, but also present in whole blood and other tissues (Supplemental Fig. 4).Unfortunately, whether these tissues (especially whole blood and retina) of mice, pig and cattle express Arrdc5 is not clear.Therefore, the conclusion that Arrdc5 in mammalian species from mice to human is germ cell specific lacks important evidence.
Author Response: New RT-PCR data has been added to Figure 2A in the revised manuscript (formerly Figure 2C) that demonstrates Arrdc5 gene expression is undetectable in both the eye and whole blood of adult mice.At this time, we do not have access to eye and whole blood of cattle or pigs so have not added to the gene expression analysis for these species.Regardless, we believe that the mouse tissue analysis is indicative of Arrdc5 expression in other species and taken together with the cattle and pig tissue analyses support the conclusion of testis specificity.

Reviewer Comment 3:
A big shortcoming is that this study only describes the infertile phenotype, although the authors state that Arrdc5 serves as a "core regulator".Without defined molecular mechanism, it is difficult to interpret the current data as "core" regulator of spermatogenesis.Similar defects have been reported in other knockout lines, is Arrdc5 functionally interacting with these proteins such as Capza3, or Ccdc62 to serve its core role?
Author Response: We agree with the reviewer that elucidating the mechanism of action for ARRDC5 in regulating spermatogenesis is a key next step.However, we feel that this exploration is best left for future studies that can focus on understanding the molecular function of ARRDC5.At present, the tools needed to define the protein interactome of ARRDC5 are not available.We have been unable to identify a commercially available primary antibody that is specific for ARRDC5 as shown in Supplemental Figure 6.Thus, to assess functional interaction between ARRDC5 and other molecules reported to have important roles in spermatogenesis such as Capza3 and Ccdc62, new reagents or mouse models will need to be generated.Again, we believe that this will be an important next step for future studies.With that said, we do appreciate the reviewer's concern about referring to ARRDC5 as a "core" regulator of spermatogenesis in the abstract.The phenotype of the knockout mouse model unequivocally shows that ARRDC5 plays an essential role in spermatogenesis which we consider to be core in its nature.However, we understand that the term core is considered differently amongst scientists and have revised the abstract to temper statements about the functional attributes of ARRDC5.
Reviewer Comment 4: SSC transplantation experiment was not necessary for this study.Only one sperm is shown in Fig. 8B and it appears that elongating spermatids were entirely missing and SSC transplantation generated a completely different phenotype.
Author Response: Although we appreciate the reviewer's viewpoint on this experiment and opinion of the outcomes, we disagree that it is not an important piece to the story.As we stated in the final section of the Results, the purpose of the SSC transplantation experiments was to "determine whether defects in sperm production of Arrdc5-/-mice are intrinsic to germ cells".
The SSC transplantation directly tests this, and we believe the outcomes demonstrate unequivocally that lack of ARRDC5 expression in germ cells is the cause of defective spermatogenesis.Even if the defective regenerated spermatogenesis originating from Arrdc5-/-SSCs in a recipient testis appears different compared to steady-state spermatogenesis in an Arrdc5-/-mouse testis, the conclusion that impaired capacity for normal spermatogenesis is intrinsic to ARRDC5-deficient germ cells is still supported.We have not made revision to the manuscript in response to this reviewer critique.
Reviewer Comment 5: Arrdc5-/-sperm fertilized nude oocytes and these zygotes developed to the blastocyst stage, these data indicate that Arrdc5-/-sperm can produce normal embryos after ICSI.However, because severe DNA damage is observed in Arrdc5-/-sperm, it will be interesting to examine whether these embryos can develop normally in vivo.
Author Response: We agree with the reviewer that determining whether any embryos generated from Arrdc5-/-sperm can develop normally in vivo is an important next step.However, we also want to point out that only 8% of embryos fertilized by Arrdc5-/-sperm advanced to blastocyst stage in vitro.This is a very low percentage, thus the chances of obtaining a pregnancy by embryo transfer after ICSI is challenging.In new experiments, we have performed two rounds of ICSI with Arrdc5-/-sperm and found that both the cleavage rate and blastocyst rate are reduced compared to wild-type sperm.Indeed, the in vitro blastocyst rate following ICSI with Arrdc5-/-sperm is only 4%.These new data have been added to Supplemental Figure 12 of the revised manuscript.Based on the in vitro development rates, obtaining pregnancies from ICSI generated embryos would be very challenging.In our opinion, this step is outside the scope of the current study and best left to a future study that can focus on the outcomes of using sperm from Arrdc5-/-males for assisted reproductive technologies.

Reviewer Comment 6:
In human, has anyone detected Arrdc5 mutation in OAT patients?
Author Response: We are not aware of evidence in peer-reviewed scientific literature that associates mutations in the Arrdc5 gene to impaired spermatogenesis in men.However, to the best of our knowledge the level of GWAS studies on human males that are diagnosed with OAT is thin.Indeed, in comparison to the numerous studies that have been conducted on men with azoospermia (obstructive and/or nonobstructive), the data generated for men with OAT is quite limited.Thus, associating mutations in genes like Arrdc5 to abnormal sperm morphogenesis in men is certainly an important element for the reproductive biology/medicine field to explore in future studies.We believe that our report of Arrdc5 expression being enriched, if not specific, for testicular germ cells in multiple mammalian species and required for sperm morphogenesis in mice will provide impetus for investigators in the fields of spermatogenesis and male fertility to explore links with DNA mutations in men, domestic animals, and wildlife.

Reviewer Comment 7:
Figure 4J shows the cleavage rate of control and Arddc5 knockout sperm, but huge variation exists in control.Cleavage rate is from 20 to 85% for control.Why?
Author Response: We appreciate the reviewer bringing up this point as the variation with wildtype sperm from this experiment is a bit outside of the norm.There are a multitude of possible reasons including biological variation or variations in standard reagents that were used for IVF.Regardless, the same conditions were applied to sperm from Arrdc5-/-and Arrdc5+/+ mice.
Although there was a wide variation in cleavage rate for Arrdc5+/+ sperm across the 6 different IVF sessions that were performed for this study, 0 cleavage stage embryos were generated with sperm from any Arrdc5-/-male.Thus, we stand by the conclusion drawn from these data that sperm of Arrdc5-/-mice are unable to fertilize intact oocytes.
Author Response: This is an interesting question raised by the reviewer but unfortunately is difficult to fully address.As we have shown, ARRDC5 deficiency leads to an array of sperm abnormalities ranging from seemingly normal heads but with a bent neck to macrosperm to very misshapen heads to morphologically malformed midpieces and multiple axonemes.Based on these observations, it is logical to postulate that microtubule cytoskeleton structure will vary across the different sperm abnormalities.If there was a singular oddity to the sperm of Arrdc5-/mice, clearly assessing normalcy of the microtubule cytoskeleton would be straightforward.However, with the array of different abnormalities for sperm from Arrdc5-/-mice, parsing out any abnormalities of microtubule cytoskeleton structure with different sperm abnormalities is almost a study unto itself.We believe that the reviewer's question is an important one and to properly address it requires deep characterization to fully assess microtubule structure in each different type of sperm abnormality observed for Arrdc5-/-mice.In our opinion, this effort is out of scope for the current study and warrants addressing in future focused studies.
Reviewers' comments: Reviewer #1 (Remarks to the Author): The authors have addressed my initial concerns and the paper should be published -congratulations on a very interesting story Reviewer #2 (Remarks to the Author): I found this a disappointing revision in that very few of the revisions I suggested (beyond typographical) appear to have been seriously considered.

Specifically
Comment 1 -the request for mechanistic data was not addressed.This on its own may be understandable.
Comment 11 -acrosome formation and the request for pre-ionophore data.Re reviewers seem to be splitting hairs here.The presence of significant acrosome defects is important and should be considered more carefully.Noting, and contrary to what is in the text, acrosome formation initiates early in spermiogenesis suggesting ARRDC5 has a role early in haploid germ cell development.It is a pity the authors did not chose to explore this more through eg the inclusion of EM data as well as testing P4 and ionophore induced acrsome function.
Comment 12 -and related to above, the grouping of stages makes even less sense that attempting to group individual stage.I assume that the purpose to the analysis was to determine where defects first appear in KOs.This should be determined at a cell type (step) level and will likely require higher resolution than light level tissue histology.As per the point above, I suggest you look closely at acrosome formation.This will in form the IVF data.
Comment 13 -the decision to remove data and claims rather than replace the EM images is curious.If these abnormalities were significant, why not replace the images?Data to address the underlying cause of motility defects is now completely missing.The arrows indicating the anterior acrosome (bottom right) and the sperm nucleus are both pointing to head-tail-coupling (neck) region.Indeed what is meant by anterior acrosome is not clear at all.Do you mean the post-acrosomal region?Given these errors, it is difficult to have confidence in any of the EM data.
Comment 16 -the revised text is more confusing than the original version.Perhaps break it into two sentence.Do the authors mean 'Discovery of evolutionary conserved genes expressed specifically in testicular germ cells with an essential role in determining sperm function is uncommon.'If yes, as per my original comments I dispute this claim.There are many germ cell-enriched gene expression models that result in male infertility.The majority are expressed in multiple species, including beyond mammals.
These comments plus the realisation from other reviewer's comments that the mouse model did not result in tagged-AARDC5 but rather GFP alone, has tempered my enthusiasm for this study.
As an aside, mouse vision cannot be accurately determined by watching behaviour in a standard mouse cage.https://pubmed.ncbi.nlm.nih.gov/28760697/

Response to Referee Comments/Critiques
We thank the reviewers for continued effort in trying to improve the merit and clarity of the manuscript.
We have tried to address remaining concerns with revisions to the text and figures or explaining why a concern has not been addressed.All revisions to the manuscript text for the most recent round of review are highlighted in blue.Revisions made from the first round of review are highlighted in yellow.Overall, we believe strongly that the major findings presented in the manuscript are novel, advancing of knowledge, and fully supported by the data; these include, 1) expression of ARRDC5 is testis enriched, if not specific, in mice, pigs, cattle, and humans (this has never before been described), 2) ARRDC5 has an essential role in sperm morphogenesis (the knockout mouse model clearly demonstrates this and a role for ARRDC5 in mammals has not been described previously), and 3) sperm generated in the absence of ARRDC5 are unable to fertilize eggs and yield embryos (the mating trials, in vitro fertilization, and intracytoplasmic sperm injection data clearly support this and these findings have novel implications in assisted reproduction, male fertility assessment, and male contraceptive development).In addition, we believe the multispecies integrated testicular transcriptome database produced in this study will be of major utility to investigators in the fields of developmental and reproductive biology.

Response to Reviewer 2
Summary Comments by the reviewer: I found this a disappointing revision in that very few of the revisions I suggested (beyond typographical) appear to have been seriously considered.
Author Response: We appreciate the reviewer's perspective that the first revision fell short of expectation but assure the reviewer that we did take the suggestions seriously.In the first round of review, the reviewer made 16 different critiques/suggestions.We responded to all with inclusion of new data, revisions to figures, and revisions to the text.Of note, 3 of the 16 critiques/suggestions made by the reviewer asked for consideration of new data; the remaining 13 of 16 critiques/suggestions were addressable by revisions to the text only.
One of the critiques (previously Comment 1) asked for assessment of protein ubiquitination.As we explained before, this analysis is not straightforward because ARRDC5 lacks the key domain of other -arrestins that are required for interaction with HECT type E3 ubiquitin ligases and therefore is unlikely to have a classical role in ubiquitination.This perspective was shared by Reviewer 1 who suggested that assessing the molecular function of ARRDC5 should be a goal of future studies.
Another request (previously Comment 11) for additional data to visualize acrosomes was "Some high magnification sperm images (stained with H&E) might be helpful".We have added these data to the manuscript as Supplemental Figure 11D.
The reviewer also asked if non-ionophore challenged Arrdc5 KO sperm stain for PNA binding.
Because ionophore treated sperm from the KO stain for PNA binding we struggled to understand how adding these data would fill a void in understanding.Thus, we did not include data on nonionophore treated sperm in the first revision.However, we have now performed PNA binding and staining with non-ionophore treated sperm and observed no differences compared to postionophore treatments (new images provided as Supplemental Figure 12 and quantitation included in revised Figure 4G).We believe that these data support a conclusion that ionophore induced reaction of whatever acrosomal components are formed with Arrdc5 KO sperm is significantly reduced compared to wild-type sperm.
The reviewer also made a statement that TEM images in Figure 6C are not informative (previously Comment 13).Because details of what was lacking in the images to render them uninformative were not provided, we struggled with how to address the comment but tried by adding labels and arrows to structures in the images that point out what we observed to be abnormal in Arrdc5 KO sperm compared to wild-type sperm.In the second revision we have added more TEM images of the Arrdc5 KO sperm as a new Supplemental Figure 14 that provides further evidence of abnormal sperm structures.Collectively, we believe that the images provide for Figure 6 and Supplementary Figure 14 provide sufficient evidence to support our conclusion that sperm generated in the absence of ARRDC5 function have an array of structural abnormalities.
Lastly, the reviewer made a comment (previously Comment 12) that data presented in Figure 5C was odd and disagreed with some of the seminiferous epithelium stage labels in images of Figure 5A.In response, we reassessed the stages in dozens of seminiferous tubule cross-sections and revised the data presentation to be what we consider more clearly interpretable by grouping stages rather than breaking them out individually.In the second revision we have again revaluated the staging in dozens of cross-sections and the outcomes are similar to what was presented in other versions of the manuscript.In the second revised version, we applied statistical analyses to the dataset which revealed no difference between genotypes.This analysis has been included in the manuscript.Although not statistically different, we believe the data are still of value to readers for gaining an appreciation of spermatogenesis in testes of Arrdc5 KO mice.
Reviewer Comment 1 -the request for mechanistic data was not addressed.This on its own may be understandable.
Author Response: In the first round of review, the reviewer suggested that we assess protein ubiquitination to provide some mechanistic data.As we responded in the initial review, assessing protein ubiquitination in Arrdc5 KO germ cells is not straightforward and lacks strong reasoning.Again, delving into mechanism of action for ARRDC5 is difficult because the molecular structure is different compared to all other known -arrestins, thus the exploration would require significant tool building and further experimentation.As such, we believe it is beyond the scope of the current study.Without knowing the protein interactome of ARRDC5, examining a role in ubiquitination is very difficult.
Reviewer Comment 11 -acrosome formation and the request for pre-ionophore data.Re reviewers seem to be splitting hairs here.The presence of significant acrosome defects is important and should be considered more carefully.Noting, and contrary to what is in the text, acrosome formation initiates early in spermiogenesis suggesting ARRDC5 has a role early in haploid germ cell development.It is a pity the authors did not chose to explore this more through eg the inclusion of EM data as well as testing P4 and ionophore induced acrsome function.
Author Response: We agree with the reviewer that deeper investigation on defective acrosome formation in Arrdc5 KO mice is important but believe this is best left for future studies.The array of different sperm deformities that are observed for Arrdc5 KO mice makes examination of a specific type challenging which we feel requires an independent study.Indeed, Arrdc5 KO mice generate sperm with a normal appearing head morphology but with a neck and/or midpiece deformity, as well as sperm with enlarged heads and sperm with severely misshapen heads.Sperm with normal appearing head morphology possess seemingly intact acrosomal structure as evidenced by imaging from PNA binding and staining (new Supplemental Figure 12).In addition, PNA binding can be observed for sperm with malformed heads, albeit with abnormal morphology.These observations suggest acrosome formation occurs at some level during morphogenesis of most, if not all, of the sperm in Arrdc5 KO mice.
There are a several possible causes of the abnormal sperm head morphology including defective chromatin compaction and improper shedding of cytoplasm.Both possibilities could cause abnormal acrosome biogenesis as collateral damage or acrosome biogenesis itself could be directly impaired.Parsing out these possibilities requires a focused independent study and, from our perspective, is beyond the scope of the current manuscript.However, we do appreciate the reviewer's interest in knowing more about acrosome biogenesis in Arrdc5 KO sperm and have characterized this aspect further.We have now conducted pre-ionophore PNA binding assessment with Arrdc5 KO sperm and characterized and quantified the observations (new data included in Supplementary Figure 12 and Figure 4H).In addition, we have repeated the postionophore treatment PNA binding assessment (new data in Figure 4H).Collectively, the outcomes of these assessments demonstrate that sperm heads form Arrdc5-/-mice possess a range of acrosome morphologies from normal with no observable differences compared to wild-type sperm to abnormal which is to be expected considering the sever head malformation observed for a major portion of epididymal sperm.Quantitative comparison of total PNA labeled sperm heads between pre-and post-ionophore treatment revealed no difference for Arrdc5 KO mice, unlike sperm from wild-type mice of which PNA binding was significantly reduced post-ionophore treatment.These new findings provide 1) better characterization of acrosome defects, and 2) further support our conclusion that the acrosome reaction capacity, even if the acrosomal vesicle is abnormally formed, for sperm generated in the absence of ARRDC5 function is impaired.
Reviewer Comment 12 -and related to above, the grouping of stages makes even less sense that attempting to group individual stage.I assume that the purpose to the analysis was to determine where defects first appear in KOs.This should be determined at a cell type (step) level and will likely require higher resolution than light level tissue histology.As per the point above, I suggest you look closely at acrosome formation.This will in form the IVF data.
Author Response: Our goal for conducting the staging analysis was to provide another piece of information for whether spermatogenesis overall is disrupted in the absence of ARRDC5 function.The analysis was not intended to determine where defects first appear or relate it to acrosome formation.Indeed, we draw no conclusions about acrosome biogenesis from the data.We agree with the reviewer that close examination at the cell type (i.e.spermatid step) level is needed to clearly understand where defects in spermiogenesis initiate but as discussed previously, feel that this depth of analysis is out of scope for the current study and properly addressing it is best left to future studies due to the complexity of sperm abnormalities that occur in Arrdc5 KO mice.Indeed, an array of different sperm deformities are observed for Arrdc5 KO mice ranging from morphologically normal heads but with bent necks to severely malformed heads.Linking which spermatids in the seminiferous epithelium to sperm that will have specific structural defects would require significant further investigation.For this reason, we do not believe examination of spermatids histologically for defining when impairments in acrosome biogenesis may initiate is plausible.To do this will require associating specific characteristics or lack thereof with propensity to yield sperm with specific deformities and this will require a focused independent study.
To address validity of the staging analysis, we have again reanalyzed cross-sections of testes from Arrdc5-/-and Arrdc5+/+ littermates for distribution of stages of the seminiferous cycle.
Regardless of whether the data are presented as all 12 stages or as groupings of stages there is still misalignment between the two genotypes.The data were generated in a blinded and unbiased manner; thus, we believe the findings are real and conclusions are supported.In the second revised version of the manuscript, we have applied statistical analysis to the dataset and found the differences between genotypes is not significant.This information has been incorporated into the manuscript.Although not significantly different, we still believe the data are important for providing readers with an overall assessment of spermatogenesis in testes of Arrdc5 KO mice.
Reviewer Comment 13 -the decision to remove data and claims rather than replace the EM images is curious.If these abnormalities were significant, why not replace the images?Data to address the underlying cause of motility defects is now completely missing.The arrows indicating the anterior acrosome (bottom right) and the sperm nucleus are both pointing to head-tail-coupling (neck) region.Indeed what is meant by anterior acrosome is not clear at all.Do you mean the post-acrosomal region?Given these errors, it is difficult to have confidence in any of the EM data.
Author Response: We appreciate the reviewer's perspective but are unsure what is being referred to as removed data.Perhaps the reviewer is referring to the revised way we are reporting data on distribution of stages of the seminiferous epithelium.If so, all the original data is still represented but grouped into broader stages.We revised the data reporting in response to the reviewer's comment that the original data presentation as all 12 stages of the seminiferous cycle was difficult to understand.We believe the staging data to be accurate and appreciate that different interpretations can be made based on differences in perspectives.However, we also feel that our interpretations of the data that a specific stage could not be applied to ~8% of seminiferous tubules cross-sections from testes of Arrdc5 KO mice and because of this distribution of cross-sections as stage groupings was misaligned compared to wild-type control mice are supported by the data.
Regarding the removal of claims, perhaps the reviewer is referring to the sentence we removed from the text in response to suggestion from the reviewer; however, that sentence was speculation and not an interpretation or conclusion drawn from the data.
Regarding why we did not replace EM images for Figure 6C, in the first round of review the reviewer asked to "please indicate the defects listed in the text in the figures" which we did in the revision by adding indications of observed abnormalities to the images.In the second revision, we have included new TEM images as Supplementary Figure 14 that we feel provide further evidence of the array of structural defects for Arrdc5 KO sperm.Overall, we believe the EM images clearly show abnormalities for sperm from Arrdc5 KO mice compared to wild-type controls and our observations of disorganized mitochondrial sheath, abnormal post-acrosomal segment, and malformed midpiece are supported.
Regarding the now complete absence of data to address the underlying cause of motility defects, we are unsure what the reviewer is referring to.In neither the initial submission nor the revised manuscript did we present data to explain the underlying cause of motility defects.To do so would require significant further experimentation as there could be an array of possibilities.We believe the data presented clearly demonstrate that sperm from Arrdc5 KO mice have a motility defect but delving into the underlying cause(s) is best left to future studies and is beyond the scope of the current manuscript.
Regarding errors in labels applied to the EM images that make it difficult to have confidence in the data, again we are unclear what the reviewer is referring to.For Figure 6C, we have not included arrows indicating sperm nuclei and the arrow pointing to the acrosome was intended to show a misshapen sperm head where there is clearly not a post-acrosomal segment.Our intent with using the term "anterior acrosome" was to describe the outer acrosomal membrane on the anterior portion of the sperm head.Perhaps this is not the appropriate description; however other studies have referred to the acrosome covering the anterior portion of the sperm head as the anterior acrosome.In the results section, we had described the observation as sperm lacking a post-acrosomal segment, not an anterior acrosome.For consistency and accuracy, we have relabeled the images in Figure 6C with post-acrosomal segment or PS in place of anterior acrosome or AA.Overall, we believe that the EM images accurately reflect abnormalities between wild-type and Arrdc5 KO sperm.
Reviewer Comment 16 -the revised text is more confusing than the original version.Perhaps break it into two sentence.Do the authors mean 'Discovery of evolutionary conserved genes expressed specifically in testicular germ cells with an essential role in determining sperm function is uncommon.'If yes, as per my original comments I dispute this claim.There are many germ cell-enriched gene expression models that result in male infertility.The majority are expressed in multiple species, including beyond mammals.
Author Response: We appreciate the reviewer's opinion on this statement and believe the concern being voiced reflects differences in professional perspectives.Our statement was, "Discovery of genes expressed specifically in testicular germ cells that have an essential intrinsic role in regulating spermatogenesis and functional translation from discovery in mice to evolutionary conservation in other mammalian species has been limited".We have not stated, nor do we intend to imply, that it is uncommon to discover genes expressed specifically in testicular germ cells with a role in sperm function.We stand by the perspective that discovery of genes with testicular germ cell specific expression that have intrinsic roles in sperm formation for mammalian species beyond the mouse has been limited.We note that germ cell-enriched and germ cellspecific do not have identical meanings.We don't disagree that there are models of genes that are germ cell-enriched that have conserved roles in sperm function.However, the infertility in many of the models is not necessarily do specifically to an intrinsic role in germ cells as expression of the gene also occurs in some male reproductive tract somatic cells.Also, examples of demonstrated functional roles in mammals beyond the mouse are indeed limited.To address the reviewer's concern, we have attempted to revise the statement for clarity.
New Reviewer Comment: These comments plus the realisation from other reviewer's comments that the mouse model did not result in tagged-AARDC5 but rather GFP alone, has tempered my enthusiasm for this study.
Author Response: The Arrdc5-eGfp model was generated to assess the testicular cell types that ARRDC5 protein is present in.This model produces a fused Arrdc5-eGfp transcript and the EGFP portion is liberated during translation.Thus, imaging for EGFP provides an accurate representation of the cell types that possess ARRDC5 protein.This approach has been used by a multitude of previous studies to accurately assess the cell types within a tissue that the protein product of a gene of interest is present in.Overall, we believe the model produced in this study is of major value and the imaging conducted with it provides important insights into the cell type expression profile for ARRDC5.

New Reviewer Comment:
As an aside, mouse vision cannot be accurately determined by watching behaviour in a standard mouse cage.https://pubmed.ncbi.nlm.nih.gov/28760697/