HIPK4 is essential for murine spermiogenesis

Essential revisions:

1) A detailed analysis of F-actin localization across the steps of spermiogenesis (both before and after step 8) is required. Notably, there is no direct proof of F-actin instability (or a phenotype) in the acroplaxome, or in the cytoplasm of spermatids, from steps 1 to 8, the time period corresponding to the expression of Hipk4 mRNA. It was only after HIPK4 expression, during spermatid elongation and nuclear condensation that an overt phenotype became apparent. While it is possible the absence of phenotype is the result of a dependence on external (Sertoli) cell forces, no evidence, or discussion, is included in the manuscript. As such, we request a detailed analysis of F-actin across the relevant steps of spermiogenesis (ideally using both phalloidin and antibody methods). Localisation at an ultrastructural level would be best, however, we appreciate this is entirely dependent on the affinity of the antibody — please include a comment on this in the response to reviewers.

We appreciate the reviewers' concern that our original manuscript had no evidence of actin dysregulation in Hipk4^{–/–} spermatids during the time period of HIPK4 expression, and we have investigated this question further by immunofluorescence microscopy. In addition to the fluorescently labeled phalloidin used for our previous studies, we have identified a commercially available anti-β-actin antibody that is an effective reagent for imaging the actin cytoskeleton in isolated spermatids. As suggested by the reviewers, we have employed these probes to analyze actin structures across the multiple steps of spermiogenesis, and our results are shown in Figure 7A (anti-β-actin antibody) and Figure 7—figure supplement 1 (Alexa Fluor 647-conjugated phalloidin). While phalloidin staining of isolated round spermatids was difficult to detect, possibly due to technical issues, we were able to achieve robust β-actin immunostaining at each step of spermatid differentiation. Using the anti-β-actin antibody, we can now observe actin defects in isolated Hipk4^{–/–} spermatids occur at steps when HIPK4 would normally be expressed. Beginning at steps 5/6, approximately half of the isolated mutant spermatids had fragmented acrosomes, which were frequently uncoupled from the actin cytoskeleton. In comparison, all isolated wild-type round spermatids had intact acrosomes that remained closely associated with the underlying actin-scaffolded acroplaxome. We note that these cytoskeletal phenotypes become manifest earlier than those observed by transmission electron microscopy of testis sections, and we interpret this difference as further evidence for their dependence on external forces. Such forces are normally exerted by Sertoli cells during spermatid elongation, and it is likely that the process used to isolate round spermatids from dissociated testes applies similar forces to the acrosome–acroplaxome complex.

We also thank the reviewers for suggesting that we conduct ultrastructural studies of actin dynamics during spermiogenesis. While we have now found that confocal microscopy is sufficient to confirm HIPK4-dependent actin functions in found spermatids, we consulted with Prof. Pablo Visconti, who has recently used 3D-STORM to image F-actin in mature sperm (Gervasi, et al. J. Cell Sci. 2018, 131, jcs215897. doi:10.1242/jcs.215897). Dr. Visconti advised us against applying 3D-STORM to differentiating spermatids, as the experiments would be technically demanding and likely require several months to a year to complete. He instead suggested that we obtain super-resolution images of wild-type and Hipk4^–/– spermatids by structured illumination microscopy (SIM). Unfortunately, those studies did not yield additional insights into how HIPK4 regulates the actin cytoskeleton (Author response image 1., and we have not pursued them further.

Author response image 1

Download asset Open asset

2) We also request an analysis/proof of the interaction between HIPK4 and actin within the acroplaxome. This could be achieved using a proximity ligation methods. We recommend using the knockout mouse as a control.

To investigate whether HIPK4 can interact with the actin cytoskeleton of spermatids, we isolated F-actin from whole testis lysates by ultracentrifugation. Strikingly, we observed that HIPK4 predominantly co-precipitates with the F-actin-containing pellet; only a small fraction of HIPK4 remains in the soluble fraction (which also contains G-actin). In addition, we found that loss of HIPK4 function specifically increases the amount of CAPZA2, an actin capping protein, that co-precipitates with F-actin. Together, these results suggest that HIPK4 exists primarily in complex with the actin cytoskeleton and/or actin-binding proteins. They also provide further evidence that HIPK4 regulates the actin cytoskeleton in spermatids.

It is important to note that our mechanistic hypothesis does not require direct interactions between HIPK4 and the actin cytoskeleton within the acroplaxome. Our immunofluorescence studies demonstrate that HIPK4 protein is distributed throughout the cytoplasm of step 4-8 spermatids (the acrosomal staining in elongated spermatids is non-specific). We also determined that HIPK4 overexpression in NIH-3T3 cells alters the phosphorylation state of several proteins that can interact with the actin cytoskeleton but not that of actin subunits themselves. Based on these findings, we favor a model in which HIPK4 directly phosphorylates one or more regulatory proteins that control actin dynamics.

While we appreciate the reviewers' suggestion of proximity ligation assays, their propensity to yield false-positives is a complicating factor (bioRxiv 2018; doi: 10.1101/411355), especially since the direct substrate(s) of HIPK4 (and therefore appropriate specificity controls) remain unknown. We share the reviewers' interest in understanding how HIPK4 regulates actin within the acroplaxome, and this is the focus of our future studies.