Mouse Y-Encoded Transcription Factor Zfy2 Is Essential for Sperm Head Remodelling and Sperm Tail Development

A previous study indicated that genetic information encoded on the mouse Y chromosome short arm (Yp) is required for efficient completion of the second meiotic division (that generates haploid round spermatids), restructuring of the sperm head, and development of the sperm tail. Using mouse models lacking a Y chromosome but with varying Yp gene complements provided by Yp chromosomal derivatives or transgenes, we recently identified the Y-encoded zinc finger transcription factors Zfy1 and Zfy2 as the Yp genes promoting the second meiotic division. Using the same mouse models we here show that Zfy2 (but not Zfy1) contributes to the restructuring of the sperm head and is required for the development of the sperm tail. The preferential involvement of Zfy2 is consistent with the presence of an additional strong spermatid-specific promotor that has been acquired by this gene. This is further supported by the fact that promotion of sperm morphogenesis is also seen in one of the two markedly Yp gene deficient models in which a Yp deletion has created a Zfy2/1 fusion gene that is driven by the strong Zfy2 spermatid-specific promotor, but encodes a protein almost identical to that encoded by Zfy1. Our results point to there being further genetic information on Yp that also has a role in restructuring the sperm head.


Introduction
The mouse Y chromosome has two copies of a gene, Zfy, which together with the closely related X chromosome linked Zfx, encode zinc finger transcription factors [1][2][3]. They derive from an autosomal precursor that was added to the mammalian sex chromosomes via the pseudoautosomal region (PAR) subsequent to the separation of the marsupial and eutherian lineages [4,5]. Postnatal expression of the mouse Y genes (Zfy1 and Zfy2) is restricted to spermatogenic cells [6][7][8]. Zfy2 encodes a much more potent transcription factor than Zfy1 [9,10] and has also acquired an additional strong spermatid-specific promoter [6] suggesting an important role during spermiogenesis. Although these mouse 'Zf' genes were identified in [1989][1990] [1,2,11] only recently have any spermatogenic functions been ascribed to them [12][13][14].
This study is a sequel to our two earlier papers documenting roles for mouse Zfy with respect to the apoptotic elimination of spermatocytes with univalent chromosomes at the first meiotic metaphase [13] and for the completion of the second meiotic division [14]. The latter paper has an extended introduction that describes the 'raison d'être' for the Zfy transgene additions to Yp gene deficient mice; this will be unfamiliar to most readers and can be accessed via the link http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt= Citation&list_uids=24967676.
In 2012 [15] we established that in X Eif2s3y OSry males (Fig 1D) in which the only Yp genes present are an X located Eif2s3y transgene and an autosomally located Sry transgene, there is a block at step 7 of the round spermatid stage of sperm development, whereas in XSxr a O and X Eif2s3y Sxr b O males (Fig 1B and 1C) there was evident sperm morphogenesis. Sxr a provides a near complete Yp gene content, whereas Sxr b has a very limited Yp gene content (Fig 1A-1C); however, Sxr b is unique in having a Zfy2/1 fusion gene that includes the strong spermatid specific promotor from Zfy2. This led us to hypothesize that this spermatid specific promotor is of paramount importance in enabling Zfy to provide an essential function during sperm morphogenesis-in normal males this function would primarily be provided by Zfy2. In the current study we show through Zfy1 and Zfy2 transgene additions to Yp gene deficient models that this is indeed the case.

Results
For our published study showing that Zfy1 and Zfy2 have an important role in promoting completion of the second meiotic division to form haploid spermatids [14] we added a minute sex chromosome (Y ÃX ) [16,17]) to the three XO models. Y ÃX (Fig 1E) comprises a complete pseudoautosomal region (PAR), a short 'X telomeric region', some repeated sequences mapping adjacent to the X centromere, and a presumed X-derived centromere [18][19][20]. Y ÃX was previously shown to enable the formation of a sex bivalent, thus circumventing the MI SAC [21]. We abbreviate the three XY ÃX models as XY ÃX Sxr a , X E Y ÃX Sxr b and X E Y ÃX Sry; the X-located Eif2s3y transgene is denoted X E . Importantly in the context of the present study, we found that the addition of Y ÃX had no effect on the frequency of haploid spermatids in the context of Sry, but it markedly increased the haploid frequency in the context of Sxr b ( Figure 3A in reference [14]). Therefore, before assessing the effects of Zfy transgene additions on sperm morphogenesis we sought to establish: (1) whether the Y ÃX addition per se had any discernible effect on spermiogenic progression by comparing X E OSry with X E Y ÃX Sry; and (2) whether the markedly increased haploid frequency associated with the Y ÃX addition in the context of Sxr b had any discernible effect on spermiogenic progression by comparing The addition of Y* X does not enhance spermatid elongation or the associated sperm morphogenesis For the purposes of this study the key spermiogenic steps are 1-12 that are present in tubule stages I-XII. Haploid round spermatids first appear at stage I as a consequence of the two meiotic divisions occurring in the preceding stage XII. During stages I-VII the round spermatids form an acrosomal cap covering one side of the nucleus. During stages VIII-XII the main features of sperm morphogenesis take place (formation of the sperm tail; restructuring and condensation of the spermatid nucleus to form the sperm head). Spermatid development in all three XO models (Fig 1B and 1D) is aberrant and delayed to varying degrees [15], so assessment of tubule stage has to be based entirely on the spermatogonial and meiotic stages (see Materials and Methods).
The spermiogenic progression in X E Y ÃX Sry males proved to be equivalent to that in X E OSry males ( Fig 3A). In epithelial stage VIII the spermatids fail to re-orientate the acrosome to face The Sxr a -derived deletion variant Sxr b has a 1.3 Mb deletion (Δ Sxr-b ) removing 6 single copy genes and creating a Zfy2/1 fusion gene spanning the deletion breakpoint ( †). The deleted gene Eif2s3y is necessary for normal spermatogonial proliferation, so an X-located Eif2s3y transgene has been added. (D) X Eif2s3y OSry. This model has only two Yp genes-the testis determinant Sry provided as an autosomally located transgene and the spermatogonial proliferation factor Eif2s3y provided as the X-located transgene. E. Y* X . This mini sex-chromosome is an X chromosome with a deletion from just proximal to Amelx to within the DXHXF34 repeat adjacent to the X centromere ( † marks the deletion breakpoint). This X chromosome derivative has a complete PAR that can pair and crossover with the PAR of XSxr a , XSxr b or X to form a 'minimal sex bivalent'. Scale bar for magnified views is 150 kb. the basal membrane, which normally precedes spermatid elongation. Moreover, the spermatids do not elongate and remodel the chromatin to form the sperm head (even by stage XII), and in both models the arrested, still round, spermatids are eliminated by stages II-IV of the following cycle. Thus the Y ÃX addition per se has no effect on spermiogenic progression.
In agreement with our previous findings, X E Sxr b O males differ markedly from X E OSry males ( Fig 3B). In X E Sxr b O males spermatids do re-orientate such that the acrosomes come to face the basal membrane and spermatid elongation and subsequent nuclear condensation do take place. However, this spermiogenic progression is delayed relative to XY controls [15]; thus re-orientation does not occur during stage VIII, and elongation and some shaping of the sperm head with associated acrosome relocation are only evident by stage XII instead of stage IX. Correctly orientated spermatids with condensed nuclei are apparent by stages II-IV instead of X-XI. A similar pattern of spermiogenic progression can be seen in X E Y ÃX Sxr b males ( Fig 3B) with elongation and condensation of the spermatid nuclei and the correct orientation of elongating spermatids being most obvious in stage XII tubules and beyond.
Addition of Zfy2 (but not Zfy1) to X E Y* X Sry males promotes substantial sperm morphogenesis The effects of the Zfy1 or Zfy2 transgene additions were assessed in X E Y ÃX Sry males in which the only Y genes present are Sry and the X-located Eif2s3y; the Zfy transgenes used were single copy and also located on the X chromosome, and we denote the resulting males as X E,Z1 Y ÃX Sry and X E,Z2 Y ÃX Sry. In agreement with our expectation, addition of Zfy1 had no discernible effect on spermiogenic progression, with the spermatids failing to re-orientate or elongate. In marked contrast, the Zfy2 addition did promote spermiogenic progression ( Fig 4A). Indeed, tubule sections of X E,Z2 Y ÃX Sry males in stages X to VII all showed elongating spermatids with condensed nuclei and a tail being formed.
We next wished to assess whether the Zfy2 addition reinstated spermiogenic progression to the same extent as that supported by the near complete Yp gene complement present in Sxr a . For this we compared spermiogenesis in X E,Z2 Y ÃX Sry males with that in XY ÃX Sxr a males. The latter have Zfy1 and Zfy2 encoded within Sxr a but, like X E,Z2 Y ÃX Sry males, lack genetic information encoded by the mouse Y long arm that is important for sperm head shaping [22][23][24][25]. We also included wild type XY controls in order to check at what tubule stage the abnormal sperm head shaping in XY ÃX Sxr a became evident. Detailed analysis of staged seminiferous epithelium revealed spermiogenic anomalies in X E,Z2 Y ÃX Sry males relative to both XY ÃX Sxr a and wild type control genotypes (Fig 4B). At stage XI, there was a marked delay in spermatid elongation and nuclear condensation, and by stages I-III the spermatids looked unhealthy with darkly stained nuclei. Silver stained testicular cell smears showed that these abnormal spermatids were nevertheless capable of developing tails (Fig 4C). We then examined epididymal tubule sections in order to assess the relative numbers of testicular sperm passing from the In X E Y* X Sry the block to spermiogenesis remains with elimination of the arrested cells once again evident by stages II-IV (see inset). (B) In X E Sxr b O, at stage VIII the spermatids have not elongated and they are randomly orientated relative to the tubule basement membrane. However, as previously reported , spermatid elongation is delayed rather than absent, and is apparent by stage XII. Nuclear condensation is also delayed as it is not evident at stage XII, but many of the elongating spermatids survive to stages II-IV at which point nuclear condensation is now evident. In X E Y* X Sxr b spermatid elongation and nuclear condensation is similarly delayed, but there appear to be fewer elongating spermatids surviving to stages II-IV [note the now evident haploid (h) as well as diploid (d) spermatids]. Scale bar is 40 μm (insets are x3 magnification). testis into the epididymis ( Fig 4D). As previously established [21], in XY ÃX Sxr a males substantial numbers of sperm reach the epididymis although these sperm have abnormally shaped heads that rarely have a hooked tip. However, no sperm were seen in epididymal sections from X E,Z2 Y ÃX Sry males. Instead, there were many round spermatids, indicating that substantial numbers of these cells slough off from the seminiferous epithelium in the Zfy2 transgenics.
Finally we assessed sperm morphogenesis by electron microscopy. Sperm heads from XY ÃX Sxr a males elongate similarly to those of wild type males except for some acrosome invagination. In contrast, X E,Z2 Y ÃX Sry males have compromised sperm head morphogenesis with signs of necrosis, acrosome invagination, incomplete elongation, and formation of cytoplasmic vacuoles (Fig 5A). Degenerating condensed spermatids were often present in large cytoplasmic inclusions within the epithelium, indicative of removal by apoptosis (not shown). However, cross sections of developing sperm tails appeared grossly normal with the classic axoneme structure composed of a central microtubule pair and nine outer doublets (9x2+2) (Fig 5B).
Clearly, although addition of the Zfy2 transgene overcomes the step 7 round spermatid arrest to allow substantial sperm morphogenesis, qualitatively and quantitatively it does not match that achieved with the near complete Yp gene complement present in XY ÃX Sxr a males.

Discussion
The starting point for the present study was our finding that in X E OSry males (Y gene complement Eif2s3y and Sry) there is a clear-cut and unique arrest of spermiogenesis at step 7 of spermatid development [15]. In the present study we showed that this arrest is also seen in X E Y ÃX Sry males that have the same Y gene complement, but also have the minute Y ÃX chromosome that enables the formation of a sex bivalent, thus avoiding MI SAC responses. The transition from spermatid step 7 to step 8 is pivotal in spermiogenesis: at step 7 the spermatids are round and randomly oriented with respect to the basal membrane, whereas at step 8 the spermatids have re-orientated so that the developing acrosomal cap is now facing the tubule periphery. This re-orientation heralds the dramatic restructuring of the round spermatids to form spermatozoa ('sperm morphogenesis'). We hypothesized that the arrest at step 7 was a consequence of a lack of Zfy2-dependent transcriptional changes, and that the expression of Zfy2 at the critical stage was dependent on the alternative, strong, Cypt-derived, spermatid-specific promotor acquired by Zfy2 during the evolution of mice (muridae) [15]. Our present results clearly demonstrate that Zfy2 supports sperm morphogenesis, while Zfy1, which lacks this promotor, does not.
The transcription factors encoded by Zfy1 and Zfy2, together with their X-linked homologue Zfx, are predicted to bind the same DNA sequence [26][27][28]. In most eutherian mammals Addition of Zfy2 (but not Zfy1) to X E Y* X Sry enables re-orientation of spermatids, together with clear head and tail morphogenesis; however the Zfy2 transgene is not as effective as Sxr a in supporting sperm morphogenesis. Hematoxylin and eosin (H & E) stained testis tubule sections. (A) By comparison with X E Y* X Sry, it is clear that the Zfy1 transgene addition has no beneficial effect on spermiogenic progression whereas the Zfy2 addition leads to the formation of correctly orientated elongated spermatids which have condensed sperm heads and some tail formation. (B) By comparison of X E,Z2 Y* X Sry with XY* X Sxr a (and XY controls) it is clear that Sxr a is more effective than the Zfy2 transgene in supporting the sperm morphogenesis. In agreement with the illustration in Fig 2, in XY males at stage XI sperm head morphogenesis has progressed to the early 'hooked tip' stage and nuclear condensation is evident. As previously reported, in XY* X Sxr a at this stage some of the sperm from the previous cycle have not yet been shed (stained dark blue); the spermatids from the new cycle are retarded with respect to elongation and nuclear condensation is not evident. In X E,Z2 Y* X Sry spermatid elongation is further retarded. Nevertheless, by stage I-III many spermatids with condensed chromatin are found in XY, XY* X Sxr a and X E,Z2 Y* X Sry. (C) Images of testicular sperm found in silver stained testicular cell smears. All three genotypes proved to have some sperm present in testicular cell smears. In XY* X Sxr a , as previously reported for epididymal sperm and testicular sperm, the sperm heads rarely have a hooked tip. In X E,Z2 Y* X Sry the developing sperm heads show limited elongation. In all three genotypes the tails were well developed. (D) Hematoxylin/eosin stained epididymal tubule sections. In XY there are abundant sperm present with the heads showing the characteristic hooked tip. In XY* X Sxr a the sperm heads rarely have a hooked tip. In X E,Z2 Y* X Sry males no sperm could be identified; instead degenerating round spermatids are found suggesting that there has been some shedding of round spermatid stages despite the addition of Zfy2. Zfx and Zfy are widely expressed, and Zfx is exempt from X dosage compensation, suggesting a constraining dosage requirement in somatic tissues [29,30]. However, in the myomorph rodent lineage, Zfx became subject to X-dosage compensation and the Zfy-encoded proteins diverged [1,29,31], and mice (muridae) ended up with two copies of Zfy with postnatal transcription restricted to spermatogenic cells [6][7][8]. We have previously discussed the evolutionary pressures that may have led to ZFY2 having a more potent transactivation domain than ZFY1, and ZFY1 frequently lacking the transactivation domain as a consequence of alternative splicing [14]. Here we consider the acquisition of the additional Cypt-derived promotor in the muridae [6,9]. It is clear from recent studies that the mouse Y chromosome gene complement has been markedly affected by a post-meiotic X-Y genomic conflict, with the round spermatid specific genes Sly and Slx being key protagonists [22,[32][33][34][35][36]. As a consequence of this conflict there has been co-amplification of Sly and Slx (50-100 copies located on the mouse Y long arm [35]) with Sly expression serving to damp down transcription of the X and Y chromosomes in spermatids [22]. Progressive reduction of Zfx and Zfy gene transcription in spermatids as the genomic conflict continued will have provided a strong selective force for the recruitment of the strong Cypt-derived spermatid-specific promotor to Zfy2, in order to maintain sperm morphogenesis.
Our original focus on the alternative Cypt-derived Zfy2 promotor was triggered by our finding of substantial sperm morphogenesis in X Eif2s3y Y ÃX Sxr b males in which a 1.3 Mb Yp deletion (Δ Sxrb -see Fig 1C) has created a Zfy2/1 fusion gene that includes the same promotor. We have developed a new PCR assay that specifically amplifies Cypt promotor dependent transcripts with exon 6 that encodes the transactivation domain (S1 Appendix and S1 Fig). Our primary interest was to see if the Cypt promotor generated comparable levels of full length transcripts from the Sxr b -located Zfy2/1 fusion gene in X Eif2s3y Y ÃX Sxr b males as compared to those from the Zfy2 transgene in X Eif2s3y,Zfy2 Y ÃX males. Quantification of the Cypt-driven transcript bands indicates that this is indeed the case as can be seen from the last two pairs of columns of Fig  6A. It is these transcripts that we contend are supporting the sperm morphogenesis evident in the two models. From Fig 6B it can be seen that the Zfy2/1 fusion gene present in XEif2-s3yY Ã XSxrb is expressed at twice the level of the XYSxrb control; furthermore, the Zfy2 transgene in X Eif2s3y,Zfy2 Y ÃX is also expressed at a substantially higher level than the endogenous Zfy2 gene of the XYSxr b and XY controls. This elevated expression of Zfy2 is to be expected in the genotypes lacking the Y long arm gene complement because they lack the repressive effect of the estimated 50-100 copies of Sly that are involved in the post-meiotic genomic conflict. In whole testis samples from XY males with a targeted knockdown of Sly transcripts there is widespread up-regulation of spermatid expressed X and Y genes and this includes Zfy2 (see [22], The PCR primers for this assay amplified Zfy2/1 and Zfy2 transcripts; only XYSxr b has both transcripts. It can be seen that the level of Cypt-Zfy2/1 transcripts in X E Y* X Sxr b is comparable to the level of Cypt-Zfy2 transgene transcripts in X E,Z2 Y* X Sry. (B) For the left panel the PCR primers are specific for Zfy2/1; the second and third genotypes lack the Zfy2/1 fusion gene so the low level signal represents 'background'. For the right panel the PCR primers are specific for Zfy2; in this case it is the fourth genotype that lacks Zfy2. The genotypes lacking the Y long arm (Yq-) have substantially higher transcript levels than the genotypes with a complete Y (Yq+).
Our finding that sperm morphogenesis in X Eif2s3y,Zfy2 Y ÃX Sry males is inferior to that in XY ÃX Sxr a males points to there being other Yp genes that contribute to sperm morphogenesis. We have generated transgenic lines that express other Yp-located genes (Ube1y1/Uba1y, Smcy/Kdm5d, Uty, Dby/Ddx3y, Usp9y, H2al2y) that may have functions during spermiogenesis. Since our current study was completed two more male germ-line expressed genes, Prssly and Teyorf1, have been mapped to Yp [36]; these are present in Sxr a (Fig 7), so they are also candidates for the improvement in sperm morphogenesis in XY ÃX Sxr a males relative to X Eif2s3y,Zfy2 Y ÃX Sry males. However, the breeding required to generate XY ÃX males with four or more transgenes is now prohibitively expensive. It is therefore encouraging that recent advances in Y gene targeting [37][38][39] promise simpler and more rapid advances in understanding mouse Yp gene functions than our onerous strategy of Yp transgene addition.
The present findings with respect to the pivotal role of Zfy2 (but not Zfy1) in enabling the transition from randomly orientated round spermatids to spermatids undergoing sperm morphogenesis, are an important addition to our recent findings documenting sometimes overlapping but nevertheless distinct roles for Zfy1 and Zfy2 during earlier stages of spermatogenesis. It will be a substantial challenge for the future to document the direct and indirect transcriptional changes triggered by these transcription factors, which underlie their distinct roles.

Animals
The mice utilized in this study had a limited Y gene complement (see below and Fig 1). Details of their production are provided in Vernet et al., 2014 [14]. All males were examined at 6 weeks of age. 1. X Eif2s3y OSry (abbreviated as X E OSry) males carry an X-located transgene encoding the spermatogonial proliferation factor Eif2s3y [40] and an autosomally-located transgene of testis determinant Sry [41,42].
2. X Eif2s3y Y ÃX Sry (abbreviated as X E Y ÃX Sry) males have the same Y gene complement as X E OSry but carry a minute X chromosome derivative (Y ÃX ) with a complete pseudoautosomal region (PAR) [18][19][20].
3. X Eif2s3y Sxr b O (abbreviated as X E Sxr b O) males have the X chromosome carrying an Eif2s3y transgene [40] together with Tp(Y)1Ct Sxr-b attached distal to the X PAR. Tp(Y)1Ct Sxr-b is an Sxr a derivative with a 1.3 Mb deletion that has removed the majority of the Yp gene complement and created a Zfy2/1 fusion gene [9,43]. 7. X Eif2s3y,Zfy1 Y ÃX Sry (abbreviated as X E,Z1 Y ÃX Sry) males are genotype 2 males to which a single copy, X chromosome-located, Zfy1-Uba1y BAC (RP24-327G6) transgene [12] has been added.
8. X Eif2s3y,Zfy2 Y ÃX Sry (abbreviated as X E,Z2 Y ÃX Sry) males are genotype 2 to which a single copy Zfy2 BAC transgene inserted by cassette mediated exchange (CME) into the Hprt locus on the X chromosome [12,13] has been added.
The mice were fed ad libitum with a standard diet and maintained in a temperature and light-controlled room (22°C, 14h light/10h dark). The protocols for animal handling and treatment procedures were reviewed and approved by the Crick Biological Research Facility Strategic Oversight Committee (BRF-SOC) and Animal Welfare and Ethical Review Body(AWERB), in accordance with the United Kingdom Animal Scientific Procedures Act 1986 (Procedure Project Licence: 80/2186).

Histology and analysis of sperm head morphology
For histology analysis, part of the testes were fixed in Bouin overnight and then stored in 70% ethanol prior to embedding in paraffin wax, sectioning at 5 μm, and staining with hematoxylin-eosin (H&E) or periodic acid Schiff and hematoxylin (PAS-H). The stages of seminiferous tubules were identified based on the composition of cells near the basal membrane according to the method described by Ahmed [45]. This was necessary because of meiotic and post-meiotic arrests present in males with limited Y gene complement, which prevented staging based on the changes of acrosome and nuclear morphology of spermatids. As a consequence, some tubule stages cannot be distinguished reliably so a range has to be given.
Sperm smears obtained from the testis were silver stained and analyzed as described previously [41].

Transmission electron microscopy
Mice were perfused, through the left ventricle [46], with ice-cold 2.5% glutaraldehyde fixative diluted in PBS. The testes were dissected, left for 1h in the fixative, and cut into small blocks that were kept at 4°C in the same fixative until embedding. Testes were post-fixed for 1h in 1% osmium tetroxide, stained 'en bloc' for 1h with 1% aqueous uranyl acetate, dehydrated with graded alcohol series and embedded in Epon. Ultrathin sections (60nm) were contrasted 7min with uranyl acetate and lead citrate and then examined using JEOL 1200EX electron microscope.

RT-PCR band quantitation for Fig 6
Band intensity was quantified using ImageJ, background subtracted and intensity normalised to Lemd1 in each sample. For further details see S1 Appendix, S1 and S2 Figs.

PCR analysis for Fig 7
Standard PCR on genomic DNA was used to map the Y chromosome genes Prssly and Teyorf1. The controls are Tspy-ps, present in the Sxrb deletion interval, and the autosomal gene Atr. For primers and annealing temperatures see S1 Table. Supporting Information S1 Appendix. Duplex RT-PCR assays.  Table. (TIF) S1 Table. List of primers used. (DOCX)