Histochem Cell Biol DOI 10.1007/s00418-017-1611-3

ORIGINAL PAPER

Subcellular localization and characterization of estrogenic pathway regulators and mediators in Atlantic salmon spermatozoal cells Kristian R. von Schalburg1,2   · Brent E. Gowen3 · Jong S. Leong1 · Eric B. Rondeau1 · William S. Davidson2 · Ben F. Koop1 

Accepted: 23 September 2017 © Springer-Verlag GmbH Germany 2017

Abstract  Much progress has been made regarding our understanding of aromatase regulation, estrogen synthesis partitioning and communication between the germinal and somatic compartments of the differentiating gonad. We now know that most of the enzymatic and signaling apparatus required for steroidogenesis is endogenously expressed within germ cells. However, less is known about the expression and localization of steroidogenic components within mature spermatozoa. We have assembled a sperm library presenting 197,015 putative transcripts. Co-expression clustering analysis revealed that 6687 genes were present at higher levels in sperm in comparison to fifteen other salmon tissue libraries. The sperm transcriptome is highly complex containing the highest proportion of unannotated genes (45%) of the tissues analyzed. Our analysis of highly expressed genes in late-stage sperm revealed dedication to tasks involving chromatin remodeling, flagellogenesis and proteolysis. In addition, using various different embedding and microscopic techniques, we examined the morphology of salmon spermatozoa and characterized expression and localization of several estrogenic regulatory and signaling Electronic supplementary material  The online version of this article (doi:10.1007/s00418-017-1611-3) contains supplementary material, which is available to authorized users. * Kristian R. von Schalburg [email protected] 1



Department of Biology, Centre for Biomedical Research, University of Victoria, Victoria, BC V8W 3N5, Canada

2



Department of Molecular Biology and Biochemistry, Simon Fraser University, 8888 University Drive, Burnaby, BC V5A 1S6, Canada

3

Department of Biology, Electron Microscopy Laboratory, University of Victoria, Victoria, BC V8W 3N5, Canada



proteins by immunohistochemistry. We provide evidence for the endogenous synthesis and localization of aromatase (CYP19A and CYP19B1) and potential mediators of estrogen [i.e., ER-alpha and soluble adenylyl cyclase (sAC)] or phosphate (i.e., CREB and FOXL2A) signaling. Partitioning of select transcripts that encode AR-beta, FSH and the LH receptor, but not AR-alpha, LH or the FSH receptor, further points to localized specificity of function in the steroidogenic circuitry of the sperm cell. These results open new avenues of investigation to further our understanding of the intra- and intercellular regulatory processes that guide sperm development and biology. Keywords  Gamete biology · Gene expression · Immunolabeling · Signal transduction · Spermatogenesis · Testis Abbreviations AR Androgen receptor (α and β) cAMP cyclic AMP CREB cAMP response element binding protein CREM cAMP response element modulator CYP19A and CYP19B Cytochrome P450 superfamily aromatase A and B DAX-1 Dosage-sensitive sex reversal, adrenal hypoplasia congenital, critical region on the X-chromosome, gene-1 DM-domain proteins Doublesex- and male abnormal-3 (MAB-3)-related transcription (DMRT) factors E2 17-β-estradiol ER Estrogen receptor

13

Vol.:(0123456789)



FOXL2 Forkhead box L2 IHC Immunohistochemistry MBM Plastic, methyl methacrylate/ butyl methacrylate MRPL4 Mitochondrial riboprotein L4 MRPS14 Mitochondrial riboprotein S14 MIS Müllerian inhibiting substance (also known as anti-Müllerian hormone or AMH) ODF Outer dense fibers PDE Phosphodiesterase PKA Protein kinase A RPL26 Cytoplasmic riboprotein L26 RPS29 Cytoplasmic riboprotein S29 sAC Soluble adenylyl cyclase TUBA α-Tubulin

Introduction Spermatogenesis is a series of coordinated and regulated processes that entails mitosis, meiosis, proliferation and specialization (reviewed in both Kimmins et al. 2004; Schulz et al. 2010). These processes transform undifferentiated diploid stem cells into uniquely specialized, differentiated haploid spermatozoa (Kimmins et al. 2004; Schulz et al. 2010; Kleene 2013). Changes in metabolism and gene expression within Sertoli and Leydig cells orchestrate spermatogenesis under the endocrine direction of cues from the hypothalamus (GnRH), pituitary (LH and FSH) and gonads (androgens, estrogens) (the HPG axis) (reviewed in both Holdcraft and Braun 2004; Schulz et al. 2010). The gonadotropins nurture germ-cell maturation through production of androgens, estrogens and progestins, that feedback negatively and positively throughout the HPG axis (Holdcraft and Braun 2004; Schulz et al. 2010). The underlying purpose of this complex interplay between the testicular soma and germ cells is production of viable, fertilization-competent male gametes to perpetuate the species. The mature spermatozoon is a highly differentiated and compartmentalized cell. It consists of the head (containing the nucleus and the paternal genome), and four components that comprise the flagella (the motile tail of the cell): the connecting piece (at the base of the sperm head); the midpiece (containing a helically arranged sheath of mitochondria), the principal piece (posterior to the midpiece and characterized by its cytoskeletal fibrous sheath) and the end piece. During development of the connecting piece, centrosomes (and their associated centrioles) act as organizing centers for the assembly of the characteristic “9 + 2” microtubular axoneme (reviewed in both Manandhar et al. 2005; Schatten and Sun 2010). The axoneme contains a central pair of singlet microtubules that are surrounded by nine outer

13

Histochem Cell Biol

doublets of microtubules that extend centrally from the connecting piece along the length of the flagellum (Inaba 2003; Turner 2003). A number of cytoskeletal structures, known as the outer dense fibers (ODFs), surround and associate with the axoneme in different configurations along the flagellar tail (Inaba 2003; Turner 2003). The axoneme and ODFs are further encased by the mitochondrial array in the midpiece, and by the fibrous sheath (FS) in the principal piece (Inaba 2003; Turner 2003). The mitochondrial array, FS, ODFs and the axoneme act as organizing centers for various regulatory complexes that generate energy, transmit signals and/or work as motors (Eddy et al. 2003; Inaba 2003; Turner 2003; Krisfalusi et al. 2006; Valsecchi et al. 2013; Mizrahi and Breitbart 2014; Vadnais et al. 2014 and refs within all). For example, different regulatory complexes that are involved in the processes that produce the high-energy fuel and regulatory molecule ATP are found in the mitochondrial (i.e., respiratory chain enzymes) and fibrous sheaths (i.e., glycolytic enzymes) (Eddy et al. 2003; Inaba 2003; Turner 2003; Krisfalusi et al. 2006; Valsecchi et al. 2013; Mizrahi and Breitbart 2014; Vadnais et al. 2014). Also partitioned among the flagellar subcellular compartments are a myriad of different signal factors, such as protein kinases and phosphatases, as well as signal inhibitors. These signal mediators utilize the ATP (and its adenine congeners and other cyclic nucleotides) to regulate and communicate the phosphorylation state of numerous downstream intracellular mediators. These signals are communicated throughout the flagella, from external cues acting at the plasma membrane, and through the mitochondria and FS, to the axoneme (Eddy et al. 2003; Inaba 2003; Turner 2003; Krisfalusi et al. 2006; Valsecchi et al. 2013; Mizrahi and Breitbart 2014; Vadnais et al. 2014 and refs within all). Some of the responses include activation (or deactivation) of enzymes, receptors, channels and other signal propagators that ultimately power sperm motion (Inaba 2003). For example, the outer-doublet microtubules of the axoneme slide against each other, powered by ATPase motors in the dynein complex that project from them, to generate flagellar motive force (Inaba 2003; Turner 2003). Collectively, these various protein complexes, located within specialized domains, provide a vast signaling and regulatory network that extends throughout the whole flagellum to orchestrate motility in a coordinated and sustained fashion. Most of our understanding about the regulation, signaling and metabolism throughout this flagellar network comes from mammalian and green algae studies (Inaba 2003; Turner 2003). Very little is known about these molecular processes in the teleost sperm. However, it is well known that local aromatase activity and estrogen (E2) synthesis and metabolism are important for viable germ cell development for each sex (Viñas and Piferrer 2008; Guiguen et al. 2010;

Histochem Cell Biol

Schulz et al. 2010; Gohin et al. 2011; von Schalburg et al. 2013; Delalande et al. 2015). Nevertheless, our understanding of the translational expression and specific localization of the functional machinery for estrogen synthesis (i.e., aromatases) and action [i.e. estrogen receptors (ERs)] within the differentiated male gamete is unknown. Furthermore, little is understood about the intracellular mediators of aromatase action or the molecular components that transduce estrogen signals within the teleost sperm. We, therefore, explored the endogenous synthesis of known aromatase regulators (DAX-1, FOXL2A, MIS) with respect to the expression of aromatase (CYP19A and CYP19B1); and various potential mediators of estrogen (ERα and sAC) and/or phosphate signaling (CREB and FOXL2A) within the ejaculated salmon sperm cell by immunohistochemistry (IHC). Analysis of the localization and relative amounts of each antigen present in specific compartments throughout the sperm cell was enabled by immunogold labeling and transmission electron microscopy (TEM). We provide for the first time evidence for the functional partitioning of some factors (e.g., CYP19A and ERα) and clear distinctions of localized abundance for others (e.g. CREB and sAC) within the salmon sperm cell body. The potential function of these regulators, expressed within their respective compartments, is discussed. To assess its particular gene expression signature, we also compared the transcriptome of the ejaculated sperm to those of the testis and other somatic tissue libraries by co-expression clustering analysis. This analysis revealed that 6687 genes were present at higher levels in the sperm in comparison to fifteen other salmon tissue libraries. We learned that the sperm transcriptome is highly complex containing the highest proportion of unannotated genes (45%) of the tissues analyzed. Different classes of genes involved in chromatin organization, flagellar construction and for proteolytic activity were highly expressed, capturing a unique stage of sperm expression. We also wondered if it was possible for limited protein synthesis to occur within terminally differentiated sperm, conventionally considered transcriptionally and translationally quiescent cells.

Materials and methods Animals and sampling The fish sampled in this study were part of ongoing established aquaculture or commercial industry practices and purposes. Ejaculated Atlantic salmon sperm samples were collected by applying pressure on the abdomen of each of five fish (Marine Harvest Canada). The initial ejaculate was discarded and the external urogenital pore wiped dry to avoid contamination from urine, feces, or blood in

subsequent collections. Six to seven mL of semen were collected from each individual. The milt was collected into plastic cups, transferred to plastic bags, to which oxygen was added, and then sealed and stored on ice for transport to the University of Victoria. Small pieces of testis were rapidly dissected from six 22-month-old Atlantic salmon individuals (Grieg Seafoods B.C. Ltd.), placed into fixative, and stored on ice for transport to the University of Victoria. Upon arrival, approximately 1.0 mL of each collected ejaculate was placed into fixative, left for 1 h at room temperature, and then placed into the fridge overnight. All samples were fixed in Karnovsky’s 0.1 M cacodylate buffered glutaraldehyde (3%) and formaldehyde (3%). RNA‑seq and assembly The remaining sperm was centrifuged 7000 rpm for 15 min, the supernatant removed and the spermatozoa immediately frozen. Total RNA was extracted in TRIzol reagent (Life Technologies, Carlsbad, CA, USA), including 0.9–2.0 mm stainless steel beads (Next Advance, Averill Park, NY, USA), by mixer-mill homogenization (Retsch, Newtown, PA, USA), followed by DNase I-treatment and spin-column purification using RNeasy Mini kits (Qiagen, Valencia, CA, USA). Each RNA sample was then quantified and qualitychecked by spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). Approximately 12 μg of total RNA was sent to Beijing Genomics Institute (Tai Po, Hong Kong) for Illumina sequencing and the transcriptomic library assembled as described (http://www.ncbi.nlm.nih.gov/ nuccore/692751866). Co‑expression clustering and statistical analysis Sixteen Atlantic salmon tissue libraries were quantified using RNA sequencing. 74,602 putative transcripts were finalized from a reference assembly (TSA: GBRB01000001 to GBRB01074602). Fragments per kilobase million (FPKM) values were calculated (CLC Genomics Workbench v6.5) and only expressed transcripts were considered (43,775, FPKM ≥ 1.0 in at least one library). Furthermore, only those transcripts that showed expression ≥ three standard deviations from the mean were used. In total, 31,560 putative transcripts were examined by ­log2 transformation of expression values (FPKM + 1). The transcripts were clustered using the hclust function in R (http://www.R-project. org/). Clustering was performed using Ward’s method and Pearson’s correlation for distance. A heatmap was generated using heatmap.2 from the gplot library in R using scaled expression.

13



Reverse transcription and cDNA amplification The remaining RNA was converted into cDNA in 25 μL reactions with 250–500 ng total RNA using oligo(dT) 15 (Promega, Madison, WA, USA) or random hexamers (IDT, Integrated DNA Technologies, Coralville, IA, USA) and ProtoScript II reverse transcriptase according to the manufacturer’s instructions (New England Biolabs, Ipswich, MA, USA). The reactions were incubated at 42 °C for 150 min and the transcriptase heat-inactivated at 70 °C for 15 min. The primer sets for each target gene were designed using Primer3 software (http://frodo.wi.mit.edu) and synthesized by IDT. The accession numbers (http://www.ncbi.nlm.nih. gov), forward and reverse primer oligonucleotides and product sizes for each gene are presented in Online Resource 1. Approximately 200–400 ng of cDNA was used in 25 μL PCR reactions containing 0.5 U Go Taq HotStart polymerase (Promega), 1 × Flexi buffer, 1.5 mM M ­ gCl2, 0.2 mM each dNTP and 100 nM of each gene-specific primer. Each PCR was carried out under the following cycling parameters: 95 °C for 5 min, then 35 cycles of 95 °C for 30 s, 55 °C (or 53 °C) for 30 s and 72 °C for 1 min, with a long final extension of 10 min, using a Peltier Thermal Cycler PTC-225 (MJ Research, Watertown, MA, USA). Amplicons for each gene (10 μL) were separated by electrophoresis on 1.0% agarose gels and images were stored using an UVP GelDoc-It documentation system (Ultraviolet Products, Upland, CA, USA). To ensure the RNA/cDNA used in PCRs was free of somatic cell contamination, a new sample of ejaculate was collected. One mL of the sample was centrifuged at 2000 g 10 min and the upper portion removed to new tubes. Other small portions (400 μL) of ejaculate were exposed to somatic cell hypotonic lysis buffer (Mao et al. 2013). Further centrifugations were undertaken to remove seminal fluid or lysis buffer by washes in ­ddH2O. RNA from the treated cells was extracted and cDNA made exactly as outlined above. Processing and embedding All samples were processed at room temperature with gentle agitation on a rotor. Pieces of the testes were processed for light microscopy immunolabeling into methyl methacrylate/butyl methacrylate (MBM) and for TEM morphology into Epon exactly as described in von Schalburg et al. (2013). Aliquots of the sperm samples were processed into Epon for morphology and into Lowicryl HM20 for TEM immunolabeling. Lowicryl HM20 processing After graded ethanol dehydration, the sperm samples were incubated in 100% ethanol/HM20 (1:1), 100% ethanol/ HM20 (1:3), and then pure HM20 for 60 min each step.

13

Histochem Cell Biol

The samples were then incubated in new pure HM20 overnight. The next day, the samples were placed into gelatin capsules filled with new HM20. The gelatin capsules were covered by their top half and placed into a Pelco UVC2 UV Cryo Chamber maintained at −20 °C with dry ice. The UV lamp was then used to polymerise the plastic for 24 h at −20 °C. Negative staining of whole sperm Salmon sperm (stored in fixative at 4 °C) was placed onto 200 mesh copper grids containing a carbon-coated formvar film and negative stained in 2% aqueous phosphotungstic acid prior to imaging within the TEM. Purity of ejaculate Epon blocks of ejaculate were serially cut into large 0.5 μm sections and collected using a Diatome histo Jumbo diamond knife (Diatome, Biel, Switzerland). After Richardson staining (0.5% Azure II and 0.5% Methylene Blue in 0.5% sodium borate), sections were photographed using the 40× objective of the light microscope. Primaries and secondaries used in LM and TEM In addition to primaries previously reported: DAX-1, FOXL2A, MIS, CYP19A and CYP19B1 (von Schalburg et al. 2013, 2014), we examined antibodies for CREB/cAMP response element modulator (CREM) (ab5803), 17-betaestradiol (E2) (ab20626), and cytoplasmic riboproteins L26 (RPL26) (ab157111) and S29 (RPS29) (ab56224); and mitochondrial riboproteins L4 (MRPL4) (ab139650) and S14 (MRPS14) (ab151118) purchased from Abcam (Cambridge, MA, USA). We also purchased alpha-tubulin (TUBA) (sc32293) antibody from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA) and anti-5.8S rRNA mouse monoclonal IgG (Y10b) from Thermo Fisher Scientific (MA1-16628) (Waltham, MA, USA). We obtained anti-trout ERα and sAC from James Nagler and Martin Tresguerres, respectively. Complete information regarding each primary used in this study is provided in Online Resource 2. Fluorescent secondaries used for light microscopy (LM) included: AlexaFluor 568 goat anti-rabbit, goat anti-mouse, or donkey anti-goat diluted 1:100 in PBS-Ovalbumin. Gold conjugated secondaries used for TEM included: 18  nm colloidal Gold-AffinitiPure goat anti-rabbit, goat antimouse or donkey anti-goat IgG (H + L) (Jackson Immuno Research Inc., West Grove, PA, USA) diluted 1:100 in PBS-Ovalbumin.

Histochem Cell Biol

Light microscopy immunolabeling Sectioning and immunolabeling procedures used were exactly as detailed in von Schalburg et al. (2013, 2014). TEM morphology Sectioning and imaging procedures used are as detailed in von Schalburg et al. (2013), with the modification that some images were collected on a Jeol JEM 1400 TEM equipped with a Gatan SC-1000 digital camera. HM20 TEM immunolabeling Ultrathin sections were cut and mounted onto carbon and formvar coated nickel grids (Electron Microscopy Sciences, Hatfield, PA, USA). Sections were incubated with 1% (w/v) ovalbumin (Sigma-Aldrich, Oakville, ON, Canada) in PBS for 10 min, with primary antiserum diluted in PBS/ovalbumin for 1 h, washed three times (5 min each time) with PBS/ovalbumin, incubated with the secondary antibody diluted in PBS/ovalbumin for 1 h, washed three times (5 min each time) with PBS/ovalbumin and then washed four times for 1 min with deionized ­H2O. Prior to viewing, sections were stained with 5% (w/v) uranyl acetate in 50% (v/v) ethanol for 10 min, washed four times with deionized H ­ 2O, stained with 5% (w/v) lead citrate for 1 min, washed four times with deionized ­H2O and then air-dried. Immunolabeling controls To verify the specificity of the primaries, MBM- and HM20-embedded sections were incubated with PBS/ovalbumin (no antibody), preimmune serum (when available), or pre-adsorbed antiserum (where antigen available). To verify the specificity of the secondaries, they were tested using dilution tests within the protocols. In addition, the species-specific secondaries were used in tests with other species’ primaries.

Results Co‑expression clustering Sixteen salmon transcriptomic libraries were examined by co-expression clustering analysis. Each library was found to present a unique signature with each transcript cluster representing a higher than normal level of expression of one tissue in comparison to the other tissue libraries.

Distinct transcript cluster patterns for each library are shown in a heatmap (Fig. 1) and a statistical summary of the clustering analysis is provided in Table 1. Salmon sperm is shown to be a highly complex tissue with respect to the number of transcripts present. Indeed, the sperm has the second highest complexity observed in terms of the number of different transcripts in a tissue (6687), other than for the brain (7949) (Table 1). Interestingly, the ejaculated sperm expresses almost twice as many transcripts as does the testis (3556). Another attribute of the sperm library is that it contains the highest percentage of unannotated genes of all the transcriptomes analyzed (Table 1). This may partially be explained by the multifaceted roles of untranslated RNAs that are employed in the storage, stability, transport and processing of mRNA in sperm cells (Chennathukuzhi et al. 2003; Kotaja and Sassone-Corsi 2007; Krawetz et al. 2011; Sendler et al. 2013). Transcript sorting analysis The most significant transcript cluster in testis and sperm were sorted by gene expression. The top 100 genes are presented in Online Resource 3 and 4. Many of these genes may simply be remnants of transcription preceding chromatin condensation. However, many of these transcripts encode proteins that are important in the construction of various substructures within the flagella (see below). We, therefore, think that the various gene classes that are expressed do reflect dedication to specific tasks required during the last stages of sperm differentiation and compare these functions to those displayed by the testis transcriptome. We note high expression of core histone (H2a, H2b, H4) and linker (H1) transcripts in sperm, compared to histone variants in the testis (H2a.V, H2a.Z and H1.5), that point to ongoing remodeling of the sperm chromatin. We also observe high levels of other different participants in these processes that contrast between the sperm (e.g.; structural maintenance of chromosomes protein 2, Jade-1, high mobility group protein B1) (Online Resource 3) and testis (e.g., structural maintenance of chromosomes protein 3, methyl-CpG-binding protein 3, high mobility group protein T) (Online Resource 4). We assume that potential reorganization of the protein composition of the chromatin reflects restructuring of the paternal genome during late stages of sperm development. The expression of various different tubulins, dyneins and intraflagellar transport proteins indicate our capture of transcripts that are expressed in spermatogenic tubules within the testis during construction of spermatid flagellar tails (Online Resource 4). This observation is confirmed by the expression of other transcripts important for assembly of the axoneme, such as dynein intermediate chain-1, assembly factor-3 and light chain-1, with FPKMs ranging from 61.0

13



Histochem Cell Biol

to 79.0 (data not shown). Indeed, in the spermatozoa, many of the highly expressed transcripts we observe would be dedicated to functions required for maintaining the structure of the tail (i.e., tektin-3, ninein, ropporin-1) and ultimately for powering its motion, such as the various different phosphatases, kinases and ATPase components shown in Online Resource 3. Another interesting finding in the sperm library was the high expression of transcripts that encode hydrolytic enzymes (i.e., many diverse proteases and lipases), none of which were observed in the top 100 genes examined in the testis. Intriguingly, we also note expression of some unique

13

Skin

Muscle

Eye

Brain

Spleen

Head Kidney

Kidney

Nose

Gill

Pyl. caecum

Gut

Liver

Heart

Testis

Ovary

Sperm

Fig. 1  Co-expression clustering. Gene expression from sixteen tissue libraries in Atlantic salmon, resulting in fourteen distinct clusters (height = 13). The majority of co-expression clusters are represented in a single-tissue library. Darker colors represent higher expression

receptor transcripts (e.g. nattectin, syndecan-2, B cell receptor CD22) whose products could adorn the cell surface of the sperm cell and function as cell recognition motifs. Factors required for germ cell activities, somatic communication and/or post‑fertilization development Various members of different subsets of RNAs have been identified that are delivered by the mammalian sperm to the oocyte. Specific sperm-borne regulatory RNAs (i.e., miRNAs) and coding mRNAs (i.e., wnt4) may be important

Histochem Cell Biol Table 1  Statistical summary of co-expression clustering analysis of various salmon tissue libraries

Cluster library

A B C D E F G H I J K L M N Total

Tissue(s) represented

Eye Skin Gut Nose Heart, liver Testis Pyloric caecum Sperm Brain Muscle Kidney Head kidney, spleen Ovary Gill

for early zygotic or embryonic developmental functions (Krawetz et al. 2011; Sendler et al. 2013; Fang et al. 2014; Johnson et al. 2015). We note that transcripts that encode various different classes of differentiation/proliferation or cell fate/patterning factors, such as BMP, GDF, SOX and WNT family members, are present in the salmon sperm transcriptome (data not shown). Important in terms of the potential function of some transforming growth factor-beta (TGF-β) family members, type-2 receptor transcripts for TGF-β (GenBank accession no. GBRB01070335), BMP (accession no. GBRB01050099), activin (accession no. GBRB01032279) and MIS (accession no. GBRB01060236) are expressed within the salmon sperm. Intriguing from an endocrine perspective, we found the endogenous germ cell expression of transcripts for inhibin-alpha (GenBank acc. no. GBRB01059889), follistatin-1 (acc. no. GBRB01021308), somatostatin (acc. no. GBRB01069268), and the gonadotropin-alpha-2 (acc. no. GBRB01015418), -beta-1 (acc. no. GBRB01057447), and thyrotropin-beta (acc. no. GBRB01004993) subunits, as well as the receptors for GnRH-2, insulin-like growth factor and progesterone. If the gonadotropin-alpha, beta-1 and thyrotropin-beta subunits are translated, the heterodimers FSH and TSH could form within the germ cells. LH, FSH and TSH are important peripheral (reviewed in both Kumar and Trant 2001; Schulz et al. 2010) and local (von Schalburg et al. 2005 and refs. therein) regulators of gonadal steroidogenesis and gametogenesis. Interestingly, we did not find transcripts for the gonadotropin subunit beta-2 (the LH-beta subunit), or for the receptors of FSH or thyrotropin. Only the LH receptor was

Total unigenes

1382 813 769 1148 677 3556 523 6687 7949 778 763 958 4356 1201 31,560

Annotated SwissProt (1e-5)

Percent Annotated (%)

Unannotated (%)

886 493 590 805 473 2372 412 3674 5742 637 568 681 3671 930 21,934

64.1 60.6 76.7 70.1 69.9 66.7 78.8 54.9 72.2 81.9 74.4 71.1 84.3 77.4 69.5

35.9 39.4 23.3 29.9 30.1 33.3 21.2 45.1 27.8 18.1 25.6 28.9 15.7 22.6 30.5

detected in the sperm cells (Acc. No. GBRB01024932). However, the LH-beta subunit and FSH and thyrotropin receptor transcripts are present in the testis library. These results verify that select transcripts partition to the germ cell, serving as examples of localized specificity of function. Ribosomal RNA transcripts Strong bands representing the 18S and 28S rRNA transcripts, that form the backbone of the small and large subunits of cytoplasmic ribosomes, respectively, are typically observed in examination of total RNA extracted from salmon tissues (Fig. 2a). We did not detect bands for the 18S and 28S cytoplasmic rRNA transcripts in our inspection of total RNA extracted from mature sperm (Fig. 2b). Smaller weak bands were evident in the sperm extract (~ 450–600 nts in length), possibly indicating mitochondrial rRNA (12S and 16S rRNA) or other relatively abundant RNA transcripts. However, we demonstrate by PCRs that a small amount of these cytoplasmic rRNA transcripts, as well as their mitochondrial counterparts, are present within the mature salmon sperm (Fig. 2c). We also explored the expression of various genes for use as positive or negative controls in this study. Interestingly, we show that the genes that encode the androgen receptor, AR-β, and the DMRT-1 (but not AR-α), are also expressed in the sperm transcriptome (Fig. 2c). The RNA/cDNA used as template for RT-PCRs is shown to be free of genomic DNA contamination by lack of an arα product and length of the dmrt1 transcript that excludes intron between primers.

13



Histochem Cell Biol

Purity of ejaculate

a

b

The number of cells were tabulated from the 0.5 um serial section images using the Count Tool of Adobe Photoshop (Version 12.1). A few white (2) and red (4) blood cells were detected among thousands (5361) of spermatozoa (total 0.1%). No other residual cell types were identified. Immunolabeling quantification

c

Fig. 2  Total RNA and reverse-transcriptase PCR validation of rRNA transcript expression in ejaculated sperm. a Total RNA from ovary. b Total RNA from sperm. Equivalent amounts of each total RNA was loaded and fractionated on an agarose gel stained with ethidium bromide. c RT-PCR validation for expression of 12S, 16S, 18S and 28S rRNA transcripts in salmon sperm. Negative reverse-transcriptase sperm RNA amplification controls were run in parallel for each PCR experiment

The 18S and 28S rRNA transcripts are the most abundant RNA species, but in mature human spermatozoa the 28S rRNA transcript is fragmented by a mechanism that is not yet fully characterized (Cappallo-Obermann et al. 2011; Johnson et al. 2011). Sites of preferential cleavage seem to be located in internal positions within the structure assumed by the RNA within the ribosome (Johnson et al. 2011 and refs therein). The lack of complete rRNA transcripts, and the absence of specific subregions within the 28S rRNA, indicates a mechanism is active in mammalian sperm that abolishes the potential for translation on cytoplasmic ribosomes (Johnson et al. 2011). A similar fragmentation activity does not appear to be present in salmon sperm. The amplification product of the salmon 28S rRNA we present is between nucleotides 882–1514 of the salmon 28S rRNA transcript (accession no. GEGY01054965). This product bridges an equivalent position within the 28S transcript that was completely absent in the human (Johnson et al. 2011). Furthermore, if similar nucleolytic activity was active in salmon sperm, the complete assembling of each rRNA transcript without sequence omissions of specific internal regions would not have been possible.

13

Samples of salmon testes and ejaculated sperm were processed into Epon for morphological analysis by TEM imaging. For detection and immunolocalization of proteins of interest, testes were processed for LM immunolabeling into MBM, and spermatozoa into Lowicryl HM20 for TEM immunolabeling. One purpose of this study was to determine the expression, location and relative amount of each protein of interest presented by the ejaculated sperm cell. Many factors need to be considered in preparation for a TEM study, such as tissue fixation type, antigen stability, primary labeling efficiencies, and access to and visualization of antibody binding sites. Some antigens are extremely sensitive to fixation and dehydration when embedding into plastic or paraffin, or their tertiary structure or accessibility changes within the tissue under the various processing conditions (McDonald 1994; Skepper 2000). Once the shape of the antigen is changed or it is tightly bound by fixative, the antibody may no longer bind, even if strong identity of the paratope with its respective epitope exists. The harsh fixation and high temperature conditions used for Epon embedding results in excellent morphological preservation at the cost of many antigens not intact enough or accessible for labeling by immunoglobulins. The less harsh fixation and cooler polymerization conditions used for MBM and HM20 embedding results in poorer morphological preservation, but causes less damage to many antigens, allowing more efficient antibody labeling on these plastics (McDonald 1994; Skepper 2000). Testis morphology and LM imaging of MBM‑embedded sections The general morphology of the salmon testis is presented in Fig. 3. During spermatogenesis, the germinal cells divide and differentiate within cysts formed by Sertoli cell cytoplasmic extensions (Schulz et al. 2010). The LM image shows cysts of spermatogonia and spermatocytes with maturing spermatids filling the space within the spermatogenic tubular compartment (Fig. 3). All the cells within a cyst develop more or less at the same rate, while the various cysts within the testis develop at different rates (Schulz et al. 2010).

Histochem Cell Biol

Sperm morphology and TEM imaging of HM20‑embedded sections

Fig. 3  A LM image of a Richardson-stained Epon-embedded 0.5 um section showing the general morphology of the testis. The image shows cysts of spermatogonia (g) and primary/secondary spermatocytes with mature spermatids (sp) filling the space within the testis. All the cells within a cyst develop more or less at the same rate, while the various cysts within the testis develop at different rates

To determine if our antibodies of interest were viable and efficacious, we tested each of them on salmon testis embedded in MBM and by LM visualization. For the MBM immunolabeling, the MBM is dissolved away after sectioning and so the whole section depth is available for the primaries. That is the primary reason we chose to do our initial screening by light microscopy. Representative images of MIS, MRPL4 and TUBA immunolabeling of MBM-embedded testes sections are shown in Fig. 4. The MIS labeling shows good spermatogonial cell labeling with some “speckled” labeling within the spermatids (Fig. 4a). The MRPL4 labeling shows labeling of the spermatids while the spermatogonia appear unlabeled (Fig. 4b). The TUBA labeling shows good labeling of both the germinal cells and mature spermatids (Fig.  4c). The results for each of the antibodies examined are provided in Online Resource 5. Interestingly, in our initial labeling of MBM-embedded testes sections at the LM level, we found DAX-1 immunolabeling in the vicinity of the released spermatids. However, higher resolution TEM immunolabeling and imaging showed that the “speckled” labeling of DAX-1 in the MBM-embedded testis was within cellular debris and not directly on sperm-related cells (Online Resource 6). It is unclear if the source of the cellular debris that labels with this antibody is somatic (i.e., of Sertoli cell origin) or the product of spermatogenic cells from an earlier developmental stage.

Sperm samples were also negatively stained (for overall morphology), processed into Epon (for high resolution morphology), and into Lowicryl HM20 for TEM immunolabeling. We present the general morphology of the salmon spermatozoa in Fig. 5. The sperm head, midpiece and tail are shown in detail (Fig. 5a–c). As shown in other images, the midpiece completely surrounds the proximal region of the sperm tail, somewhat like a barrel (see Fig. 6e; Fig. 7b, d). The proximity to and orientation of the midpiece with the base of the flagella it ensheaths indicates how the longitudinally aligned mitochondria within it can interface with flagellar substructures (Fig. 5c, e). Representative images of TEM imaging of CYP19B1, RPS29, RPL26 and TUBA immunolabeling in different compartments of HM20-embedded sperm sections are shown in Fig. 6. The HM20-embedded sections contained thousands of separate sperm cells sectioned at various angles. After extensive observation of a large number of sectioned and labeled sperm compartments, a labeling pattern and the relative amounts bound within each sperm region was determined. The complete labeling results for each antibody is provided in Online Resource 7 and we present an overview of the amount of labeling found within each compartment of the sperm cells for each antibody (Fig. 8). The sperm image was derived from Fig. 5c. As an example of the specificity of our TEM-labeling protocol, we show longitudinal and cross-sections of sperm labeled with sAC or of preincubations of sAC primary with its antigen (Fig. 7). Clear labeling of sperm head and midpiece sections by sAC are demonstrated (Fig. 7a, b), with sAC labeling greatly reduced in controls (Fig. 7c, d). Overview of sperm immunolabeling The association of the endogenous expression of much of the steroidogenic machinery with fish male germ cells has been well established (Viñas and Piferrer 2008; von Schalburg et al. 2013; Caulier et al. 2015; Delalande et al. 2015 and refs within each). However, to the best of our knowledge, this is the first time that antibodies to most of the proteins examined here have been used in IHC studies on fish spermatozoa. Our work indicates that CYP19A and CYP19B1 in the salmon sperm have different functional assignments with localization of one aromatase restricted from the midpiece (CYP19A) and the other found throughout the sperm cell (CYP19B1) (Fig. 8). ERα appears to be localized predominantly in the sperm head and tail. Interestingly, we observe light labeling of E2 within each compartment of the sperm cell (Fig. 8).

13



Histochem Cell Biol

Fig. 4  Representative fluorescent and phase-contrast images showing labeling of MBM-embedded testes sections. a MIS. b MRPL4. c TUBA. g spermatogonia and primary/secondary spermatocytes, sp spermatids, Scale bar 50.0 ums for all images

In our overview of the sperm labeling results, small differences in labeling were noted for CYP19A (Fig. 8). Antibodies LVDQKRRGLREADKLDHIN and ELHNSDLQNLRVLESFIN were both generated from CYP19A, but the primaries were generated to different regions of the molecule (von Schalburg et al. 2013). One possible reason for the small differences in labeling between these two primaries is that the two different regions of CYP19A may react differently to the fixation and embedding protocols.

13

We previously immunolocalized DAX-1 and MIS, but not FOXL2A, in one-year-old male germ cells (von Schalburg et al. 2013). It is well established that these proteins are intimately involved in regulating the expression of aromatase, at least in the somatic layer (reviewed by Guiguen et al. 2010; Kikuchi and Hamaguchi 2013). We show here that each of these proteins was expressed in association with twentytwo-month-old testicular germ cells (Online Resource 5). However, we demonstrate that DAX-1 is not expressed

Histochem Cell Biol

endogenously within either spermatids (Online Resource 6) or spermatozoa (Fig. 8) by TEM. However, antibodies to FOXL2A and MIS lightly label most sperm compartments (Fig. 8). We attempted to detect ribosomes in the sperm using antibodies to protein components of both the large and small subunits of both mitochondrial and cytoplasmic ribosomes. Interestingly, we found by TEM that RPS29, RPL26 and MRPL4 were immunolocalized primarily to the sperm tail (Figs. 6b, c, 8). Surprisingly, MRPL4 did not localize with mitochondria, but RPL26 labeling was found within the midpiece (Figs. 6e, 8). MRPS14 did not successfully label the testis by LM and was, therefore, not tested on the sperm by TEM. To demonstrate immunolocalization of potential functional ribosomes, we used a commercial antibody generated against whole 5.8S rRNA (Y10b). Y10b recognizes epitopes within both the small and large subunits of cytoplasmic ribosomes. Y10b has been demonstrated to discriminate between whole ribosomes and rRNA among various different organelles (Garden et al. 1994). The binding of Y10b does appear to be specific since the cytoplasm of the cystic spermatogonia are well labeled, whilst adjacent spermatids are only weakly labeled in LM imaging of testicular MBM-embedded sections (Fig. 9a; Online Resource 5). This is congruent with our observation that a large portion of the ribosomes are extruded with the cytoplasm from the spermatids during their release into the tubules (data not shown). Exploring further using TEM, we found strong labeling of the ribosome-rich cytoplasm of cystic spermatogonia (Fig. 10a, b) with much reduced labeling of adjacent spermatids (Fig. 10b). Within the ejaculated sperm body, each compartment was labeled by Y10b: the head, the midpiece and the tail regions (Fig. 9c–h). The specificity of the binding of Y10b was confirmed and validated by pretreatment of the sections with RNase (Fig. 10c, d), or preincubation of the Y10b primary with mammalian ribosomes (Fig. 10e, f) prior to immunolabeling that resulted in greatly reduced labeling. We found CREM (or CREB) immunolabeling throughout the flagella and in the midpiece (Fig. 8). We note that the epitope of the CREB antibody that was used is 100% homologous to most CREM isoforms, at least those of humans. The immunogen in CREB covers the residues of amino acids 126–139 in CREM (Abcam). In mammalian spermatozoa, CREM localization seems to be confined to the connecting piece (Noda et al. 2012). We detected sAC immunoreactivity in each sperm compartment, with strongest labeling in the midpiece (Fig. 8). In the midpiece and in the proximal tail, sAC appears to be associated with the mitochondria and with the membrane, respectively (Fig. 7b). Finally, we used anti-TUBA as a positive control since flagellar tails are chiefly composed of

alpha- and beta-tubulins (constituents of the microtubules). This was confirmed by heavy labeling of only the sperm flagella (Fig. 8).

Discussion Fertilization is a daunting challenge, with sperm often traversing large distances to reach the egg(s). Differences exist between the mechanisms involved in sperm internal or external fertilization and in egg chemotaxis, recognition and penetration across species. Unlike in humans, where one sperm prevails, hundreds of thousands of salmon sperms are required to independently fertilize large egg deposits. Motility of salmon sperm, which fertilize externally, may be activated by less well-characterized surface-exposed steroid systems (reviewed by Inaba 2003). Transmembrane adenylyl cyclases and diverse voltage-gated ion channels and exchangers trigger numerous signaling pathways involved in activation of sperm motility in response to multiple environmental cues (reviewed by Inaba 2003; Turner 2003). Activation of these various membrane complexes results in ion flux, cyclic nucleotide generation and kinase-activated signaling that provoke specific effectors to gate membrane channels, liberate calcium stores and phosphorylate numerous phosphoproteins to further amplify and/or convey the membrane-activated response. Production of ATP by oxidative phosphorylation and glycolysis in the mitochondria and fibrous sheath, respectively, serves as a high-energy phosphate source, as well as a signaling and regulatory molecule (Vadnais et al. 2014). Many different adenylyl cyclases, phosphatases, glycolytic enzymes, protein kinases, adenylate kinases, phosphodiesterases (and other modulators) are colocalized to specific regions along the length of the sperm tail (Eddy et al. 2003; Inaba 2003; Turner 2003; Krisfalusi et al. 2006; Vadnais et al. 2014). These complexes contribute to a signaling network that permits coordinated use of ATP and its congeners (ADP, AMP and cAMP) to regulate normal flagellar wave propagation and pattern (Vadnais et al. 2014). Indeed, it has been proposed that specific, locally generated ratios of adenine nucleotides are required to precisely regulate and complete the flagellar waveform, at least in the mammalian sperm (Vadnais et al. 2014). In the following sections, we examine the potential functional activities of each of the factors examined here and how they may incorporate into the elaborate network that fuels and propels the salmon sperm. CREM/CREB We demonstrate an association of CREM or CREB with the salmon sperm midpiece and tail (Fig. 8). In mammalian

13



13

Histochem Cell Biol

Histochem Cell Biol ◂Fig. 5  General morphology of salmon-ejaculated sperm cells. a A

low TEM magnification image of a negatively stained whole mount of a sperm cell. Scale bar 2.0  ums. b A higher TEM magnification image of a. Scale bar 500  nms. c A longitudinal TEM section of sperm cells embedded into Epon. The proximal flagellar tail region that attaches within the sperm head and that is enveloped by the midpiece is featured. Scale bar 500 nms. d Cross-sections of tails showing the “9 + 2” microtubular arrangement of the axoneme. A central pair of microtubule doublets encircled by nine outer doublets are featured. The radial spokes that project from the outer microtubular doublets toward the central pair can also be observed. The cross-sections also clearly show the membrane that surrounds the flagella. Scale bar 250 nms. e A higher magnification view of c shows the morphology of the connecting piece that houses a basal body and tail attachment structures. Elongated mitochondria and vesicles (arrows) within the midpiece are also clearly shown. Scale bar 250  nms. h sperm head, mp midpiece, b basal body, t flagellar tails (longitudinal sections), tx flagellar tails (in cross-section), m mitochondria, v vesicles

studies, CREM activity has been limited to the expression of genes important to differentiation of haploid spermatids (Chennathukuzhi et al. 2003; Kimmins et al. 2004), but not to subsequent stages of sperm development. However, in the ejaculated sperm of boar, CREM has been localized to the connecting piece (at the base of the sperm head) (Noda et al. 2012). It is unclear what structures CREM may associate with in the connecting piece. CREM was not associated with residual cytoplasm or with the mitochondria in the boar spermatozoa (Noda et al. 2012). During mammalian spermiogenesis, centrosomes and numerous associated structural and regulatory proteins reside within the connecting piece (reviewed in both Manandhar et  al. 2005; Schatten and Sun 2010). The microtubular axoneme that runs centrally through the flagella originates from the extension of the distal centriole. However, by the spermatozoal stage of development, most components of these structures have been eliminated (Manandhar et  al. 2005; Schatten and Sun 2010). Although very little is known about the centrosome-centriole complex in salmon spermatozoa, it is possible that the loss of these components is incomplete and CREM serves a function related to the microtubular-based cytoskeletal flagellar complex originally laid down by the centrosomes. CREB has also been localized to the inner membrane of mitochondria of the mammalian brain (Valsecchi et al. 2013). Whether CREB is a downstream effector of sACcAMP-PKA-mediated phosphorylation signaling that has been demonstrated for other phosphoproteins in sperm mitochondria (Mizrahi and Breitbart 2014), remains to be determined (see sAC below). Indeed, further work should be initiated to determine if CREM or CREB serve a role in transmitting phosphate signals in the mitochondria and throughout the motility regulatory network of the flagella in ejaculated sperm.

FOXL2A and MIS FOXL2A, as a transcription regulator, and MIS, as an autocrine and paracrine signal transducer, are generally considered to be factors expressed within gonadal soma. Each regulator acts upon the steroidogenic pathway, and therefore, exert control over differentiation of the gonadal soma, as well as the germ cells (Guiguen et al. 2010; Pisarska et al. 2010; Kikuchi and Hamaguchi 2013). It also has been demonstrated that both FOXL2A and MIS (von Schalburg et al. 2013; Caulier et al. 2015; this paper) are endogenously synthesized within germ cells, adding another layer of complexity to their activity. In mammals, MIS expression has been correlated with rapid expansion of the granulosa cell layer, similar to Sertoli cell proliferation observed during embryonic development (Ingraham et  al. 2000). Coordination between the activities of FOXL2 and that of MIS (and other regulators) with granulosa cell differentiation and follicle maturation exists (Pisarska et al. 2010). In salmon, previous work has also demonstrated cell lineage-defining activities for both FOXL2A and MIS in developing organs that extend beyond differentiation of the embryonic gonad (von Schalburg et al. 2014). The dynamic interaction that occurs between FOXL2A and SF-1 (and other modulators) at aromatase promoters (Guiguen et al. 2010; Pisarska et al. 2011; von Schalburg et al. 2013 and refs within each) indicates that FOXL2A can physically interface with other mediators. Of interest to us in the context of this paper is that FOXL2 has been demonstrated to bind ERα in at least two different nucleation configurations in complexes with a variety of partners, at least in mammalian studies (Pisarska et al. 2011). How FOXL2A might function in the mature sperm is unknown, but part of its mechanism of action might be through phosphorylation. Regulation of FOXL2 activity has been demonstrated to occur through phosphorylation of serines in its C-terminal (Pisarska et al. 2010). The role of MIS in the mature salmon sperm is less clear. MIS is a secreted glycoprotein member of the TGF-β superfamily of regulators that includes activin and inhibin. MIS influences the activity of two key steroid-metabolizing enzymes within the steroidogenic pathway: cytochrome P450 17α-hydroxylase/17, 20-lyase (CYP17A1) and aromatase (reviewed in both Ingraham et al. 2000; Kikuchi and Hamaguchi 2013). CYP17A1 is a multifunctional enzyme with lyase and hydroxylase activities that can direct the production of either estrogens or androgens. Importantly, CYP17A1 function is critical for the reorganization of mitochondria within the sperm midpiece (Liu et al. 2007). The regulatory controls that mediate the dual biosynthetic functions, and the means by which CYP17A1 acts in mitochondrial sheath formation, are not yet fully elucidated. However,

13



Histochem Cell Biol

Fig. 6  Representative TEM images of labeled HM20-embedded sections of ejaculated salmon sperm. a CYP19B1 labeling of the head, midpiece, and tails within the section. b RPS29 labeling of sperm tail

pieces. c RPL26 labeling of tail pieces. d TUBA labeling of flagella. e RPL26 labeling of a cross-section through the midpiece. Scale bar 500 nms applies to all panels

a unique squalene monooxygenase (epoxidase) activity exhibited by CYP17A1 may be important for the integrity of the sperm mitochondria (Liu et al. 2007). MIS signaling occurs through interactions between ligand-specific type-2 and one or more recruited type-1

transmembrane serine/threonine kinase receptors. The MIS ligand and its cognate type-2 receptor are found in the supporting somatic cells, but not within the germ cells, of male and female gonads of mammals and fish (Ingraham et al. 2000; Shiraishi et al. 2008 and refs within both). However,

13

Histochem Cell Biol

Fig. 7  Examination by TEM of longitudinal and cross-sections of ejaculated sperm embedded in HM20, labeled with sAC or with sAC and antigen. a Longitudinal section of sperm cell labeled for sAC. The head, midpiece, and assorted tail pieces within the section are labeled (electron dense colloidal gold dots), while the background is clean of labeling. b Cross-section of sperm midpiece labeled for sAC. The midpiece is well labeled, while the background is clean of labeling. Two pieces of embedded tails are also labeled. c Longi-

tudinal section of sperm cell labeled for sAC after the primary was preincubated with its antigen. The labeling observed in a is greatly reduced. There is one gold within the head (arrow) and a couple of golds within the background where no sperm material is present (arrows). d Cross-section of sperm midpiece labeled for sAC after the primary was preincubated with its antigen. The positive immunolabeling shown in b is absent. The cross-sections of sperm tails also show no labeling. Scale bar 500 nms applies to all panels

13



Histochem Cell Biol

Fig. 8  A cartoon overview of the results found for each immunolabeling of HM20embedded sections of ejaculated sperm with TEM. The upper-left panel is an example of the coloring that represents the amount of labeling within each compartment of the sperm cells. The head is representing light labeling (light pink). The midpiece is representing good labeling (dark pink). The tail is representing heavy labeling (red). No coloring indicates no significant immunolabeling was detected

we do provide evidence for the expression of the transcripts that encode both MIS (GenBank acc. no. GBRB01005872) and its primary receptor (acc. no. GBRB01060236) within salmon sperm cells. Whether the type-2 receptor is synthesized and interacts with MIS (Fig. 8) needs to be confirmed.

13

The portion of the protein recognized by our antibody is within the prodomain N-terminal, the section that serves to stabilize and target the processed, receptor-binding, C-terminal hormone region. Interestingly, modulation of the activities of various members of the TGF-β family through

Histochem Cell Biol

Fig. 9  Examination by LM and TEM of Y10b labeling in MBMembedded testis and HM20-embedded ejaculated sperm sections. a LM fluorescence of a thick MBM section showing strong labeling of the testis cyst cells (spermatagonia) with fine speckled labeling within the released spermatids. b The corresponding phase-contrast image of a. Scale bar 50 microns for panels a, b. c A TEM image of a HM20 section of ejaculated sperm that shows labeled sperm head, two halves of the midpiece in longitudinal section (arrows) and tail

(between the midpiece halves). d HM20 section that shows another labeled sperm head. e–g TEM images of separate HM20 sections of ejaculated sperm that show labeled tails in longitudinal sections. h HM20 section showing a labeled midpiece in cross-section. For each TEM micrograph, the electron dense colloidal gold markers delineate positive ribosome immunolocalizations. Scale bar 1.0 micron for panels c–h 

13



Histochem Cell Biol

Fig. 10  Examination of ribosome labeling with Y10b in MBMembedded testis sections by TEM. a TEM image of a well-labeled spermatogonia cytoplasm. The electron dense colloidal gold markers delineate positive ribosome immunolocalizations. b Y10b labeling of the cytoplasm of various spermatogonia as well as tails of some released spermatids (t). No labeling is observed in the background. c Sections pretreated with RNase before labeling with the Y10b primary show greatly reduced cytoplasmic labeling compared to a. d Sections pretreated with RNAse before labeling with the Y10b pri-

mary show minimal labeling of spermatogonia cytoplasm and no labeling of the spermatids. e The positive immunolabeling of the cytoplasm shown in a is significantly decreased following preincubation with ribosomes prior to labeling. f Y10b labeling is absent in cytoplasm and in longitudinal (t) or cross-sections (tx) of released spermatids after preincubation with ribosomes. Scale bar 1.0 micron applies to all panels. Abbreviations: n nuclei of spermatocytes, t spermatid tails (longitudinal sections), tx spermatid tails (in cross-sections), mx spermatid midpiece (in cross section)

interactions of heteromeric N-terminal prodomains and mature C-terminal hormone domains have been demonstrated (Haramoto et al. 2006; McIntosh et al. 2008 and refs in both). Furthermore, we previously have identified

mis mRNAs that encode the truncated prodomain alone (von Schalburg et al. 2011). It, therefore, is conceivable that MIS may influence the bioactivities of other members of the TGF-β family in the sperm cell.

13

Histochem Cell Biol

Aromatase, estrogen and ER Cytochrome P450 aromatase is the terminal enzyme in the steroidogenic pathway that catalyzes the conversion of C ­ 19 androgens to aromatic ­C18 estrogens. In teleost fish, there are at least two distinct isoforms, CYP19A and CYP19B, that are most commonly associated with the ovary and the brain, respectively. Expression of four distinct ERs have been characterized in various salmonid tissues (Nagler et al. 2007; Delalande et al. 2015). Gonadal somatic aromatase activity and estrogen synthesis and function is required for development of the teleost gamete (reviewed by Guiguen et al. 2010; Schulz et al. 2010). More recent advances indicate that E2 synthesis and regulation is not attributable to the somatic layer alone but also to the endogenous expression of both aromatases (as well as other steroidogenic enzymes) and ER mRNA/protein within the germ cells of each sex (Viñas and Piferrer, 2008; Gohin et al. 2011; von Schalburg et al. 2013; Caulier et al. 2015; Delalande et al. 2015). Compartmentalization of the synthesis of E2 reveals exquisite regulatory control and communication between the germ and somatic layers, the full details of which are still incomplete. In mammalian-ejaculated sperm, ERα, ERβ and aromatase were found predominantly concentrated in the midpiece and tail (Aquila et al. 2002, 2004; Solakidi et al. 2005). In our study, we found CYP19A excluded from the midpiece and CYP19B1 present in each sperm compartment. E2 was detected in each compartment of the ejaculated salmon sperm cell (Fig. 8). The activities of the two mammalian ERs appear to be partitioned to either the midpiece or to the tail, with some overlap in the proximal tail (Aquila et al. 2004; Solakidi et al. 2005). In the salmon sperm, ERα was localized to the head and flagella, and it would be interesting in future to determine if other ERs (i.e. ERβ) immunolocalize to the midpiece, in accord with the mammalian study by Solakidi et al. (2005). In mammals, the action of estrogens is mediated through their cognate receptors, ERα and ERβ. ERs are not only found in pools localized primarily in the cytoplasm (or nucleus) but also in the plasma membrane and mitochondria (Hammes and Levin 2007). The complete repertoire of components that constitute the ER-membrane complex, as well as all of the downstream signaling effects of membrane-bound ERs, are still incompletely deciphered. Nevertheless, membrane-bound ERs are known to associate with G-proteins, and various different kinases and growth factor receptors (and other modulators), in complexes that transmit signals that generate distinct cellular effects (Hammes and Levin 2007; Aquila et al. 2004). At the stage we examine here, we presume ERα is acting nonclassically (i.e., not in transcriptional processes) and that it is bound to the plasma membrane and other structures, as characterized in

mammalian models (Hammes and Levin 2007; Pedram et al. 2012). Of interest to us is the localization of ER within the sperm tail (Fig. 8) where it possibly engages other proteins to transduce intracellular signals. Perhaps, ERα associates with the plasma membrane and/or fibrous sheath of the tail. Residence within the proximal tail membrane would bring the ER into close proximity to the mitochondrial sheath (see FOXL2A above). Earlier work has demonstrated the potential for estrogen stimulation of adenylyl cyclase and generation of cAMP through membrane-bound ER coupling with G-proteins (reviewed in Hammes and Levin 2007). Although much is known about how membrane-derived ER signals integrate their function through nuclear ER activities in a variety of other cell types (Hammes and Levin 2007), less is understood about the crosstalk that might exist between membrane and mitochondrial or intraflagellar ER signals, particularly in differentiating spermatozoa. Our demonstration that ERα associates with structures within the sperm tail provides an interesting opportunity to determine how estrogen-activated signals integrate into the flagellar motility regulatory network. sAC The predominant source of cAMP in the mammalian sperm tail is soluble adenylyl cyclase (sAC) (Hess et al. 2005). The presence of sAC in fishes has recently been reported with detection of sAC mRNA in both rainbow trout and dogfish shark testis by RT-PCR (Tresguerres et al. 2014). Unlike the ACs that reside in the plasma membrane that are responsive to G-proteins, sAC activity is distinctly modulated by local changes in bicarbonate anions, ATP and calcium (reviewed in Valsecchi et al. 2013). In salmonids, sperm motility is initiated by cAMP in a bicarbonate-dependent manner (Morisawa and Ishida 1987; Morisawa and Morisawa 1988), suggesting the involvement of sAC. We localized sAC to the head, midpiece and flagella of salmon sperm (Fig. 8). In mammals, sAC seems restricted to the mid- and principal pieces (Hess et al. 2005). Interestingly, in sea urchin sperm, sAC was immunodetected throughout the sperm cell (Beltrán et al. 2007), similar to our work. The sAC in the sea urchin sperm head was associated with the acrosome reaction (Beltrán et al. 2007), but not in mouse (Hess et al. 2005). Teleost sperm lack an acrosome, and therefore, do not undergo the acrosome reaction (Schulz et al. 2010). We, therefore, do not know the specific role of sAC in the salmon sperm head, but speculate that the sAC found in this compartment may coordinate adenine nucleotide-mediated signals that consolidate on the midpiece to regulate various aspects of mitochondrial metabolism and signaling.

13



Certainly, the heavy labeling of sAC in the midpiece, in association with mitochondria, points to this compartment as a strong localized source of cAMP synthesis (Figs. 7b and 8). cAMP nucleotides catalyzed by sAC within mitochondrial membranes have been shown to initiate PKA phosphorylation of various mitochondrial regulatory phosphoproteins (Valsecchi et al. 2013; Mizrahi and Breitbart 2014 and refs within both). These signals have been demonstrated to regulate mitochondrial metabolism coupling the Krebs cycle with the oxidative phosphorylation (OP) machinery (Valsecchi et al. 2013; Mizrahi and Breitbart 2014 and refs within both) to generate ATP. Additional energy production is provided by glycolytic enzymes that have been localized within the FS (Eddy et al. 2003; Inaba 2003; Turner 2003; Krisfalusi et al. 2006 and refs within each). The ATP derived from OP and glycolysis serves both as fuel and as a regulatory molecule in further enzyme cascades transmitted within the flagella that result in phosphorylation of proteins that execute the bending of and dynamic movement of the sperm tail. Although not yet confirmed, it is probable that a similar mechanism of sAC-cAMP-PKA signaling exists within the salmon sperm midpiece and FS that signal and modulate motility. Ribosome assembly We explored whether the potential existed for the assembly of functional cytoplasmic ribosomes and/or mitoribosomes within the ejaculated salmon sperm. Some controversy exists regarding the presence of ribosomal RNAs, ribosome function and the impediments to translation in ejaculated sperm (Diez-Sanchez et al. 2003; Gur and Breitbart 2008; Cappallo-Obermann et al. 2011; Johnson et al. 2011). Nevertheless, ribosomes exist in association with the mitochondrial array of the midpiece (Gur and Breitbart 2008) and partially assembled cytoplasmic ribosomes have been detected in the head and connecting piece (Cappallo-Obermann et al. 2011) of ejaculated mammalian spermatozoa. Our PCRs of salmon spermatozoan cDNA and transcriptomic data validate that 18S and 28S rRNA, and 12S and 16S rRNA, important structural RNAs for each subunit of cytoplasmic and mitochondrial ribosomes, remain intact in the salmon sperm cell (Fig. 2). The IHC localization of assembled cytoplasmic ribosomes (Fig. 9e–g) with the ribosomal proteins in the flagella (Fig. 8) may indicate that fully functional ribosomes reside in this compartment. It is not clear why we did not also detect riboproteins in the sperm head (Fig. 8). These results could be indicative of similar IHC sensitivity and accessibility issues as those described for CYP19A (see results). However, the ribosomes within the sperm head may not be functional (Fig. 9c, d). The ribosomes may be incomplete due to alterations or omissions of some riboproteins that occur in the head. In the mammalian

13

Histochem Cell Biol

sperm head, ribosomes presumably exist in a nonfunctionally assembled state with fragmented rRNA (Cappallo-Obermann et al. 2011; Johnson et al. 2015). This could be part of a mechanism that prevents errant translation at late stages of sperm differentiation or in early embryonic stages. Whether both mitochondrial and cytoplasmic ribosomes are present in the salmon sperm flagella remains to be determined. However, we did detect a mitoriboprotein in the tail (MRPL4), as well as a cytoriboprotein in the midpiece (RPL26) (Figs. 6e and 8): results that indicate that each class of ribosome may exist in both regions of the mature sperm cell. However, cross-reactivity of epitopes within the riboproteins we investigated with domains of other closely related riboproteins is conceivable. In addition, sharing of riboproteins between cytoplasmic and mitochondrial ribosomes may occur. Much more work is required to prove unequivocally that intact, functional ribosomes exist in the ejaculated salmon sperm. A more comprehensive sampling of ribosomal proteins is needed. Future IHC work using pre-embedding immunolabeling procedures, with other techniques (i.e., negative stain TEM or embedding into Epon and thick-section tomography), may provide further insight for detection of ribosomes within the salmon flagella and midpiece.

Conclusion We provide evidence for the local synthesis and distribution of an ensemble of estrogenic regulators, generators and signal mediators throughout the male gamete. The endogenous expression of CYP19A and CYP19B1, and of ERα, indicate some partitioning of the regulation, synthesis and signaling of E2 (Fig. 8). The precise role of MIS upon E2 activity in the sperm cell is currently unknown. However, the TGF-β signaling pathway helps to shape the sexual fate of the gonad, primarily by modulating estrogen production, and through incompletely understood mechanisms that control germ cell proliferation (reviewed by Guiguen et al. 2010; Kikuchi and Hamaguchi 2013). Future IHC work should investigate the presence and localization of other TGF-β family members (see results), their receptors (including the MIS cognate receptor), and attempt to elucidate the binding partners and signaling mechanisms that MIS engages. Furthermore, determining if all four salmonid ERs (Nagler et al. 2007; Delalande et al. 2015) are endogenously expressed, and if they localize to distinct microdomains, would further our understanding of the organization of E2 action throughout the sperm cell body. It is likely that local production of E2 in the sperm midpiece is required for viable mitochondrial morphology and function (Chen et al. 2005; Sarkar et al. 2015 and refs within both).

Histochem Cell Biol

Since the mitochondria are longitudinally aligned with the base of the flagella (Fig. 5e), some of the various signals generated within them can be transduced to act upon regulators associated with the adjoining ODFs and axoneme. The CREB, FOXL2A, MIS and sAC expressed in the midpiece (Fig. 8), could serve as gatekeepers for the transmission of signals to the flagellar motility apparatus. For example, each of these proteins can modulate phosphoryl charge in various ways: directly (sAC), through transfer (CREB and FOXL2A) and indirectly through receptor activation (MIS). Interestingly, each of these regulators are also present in the flagella (Fig. 8), potentially indicating similar functional attributes in both compartments. The various mediators they engage may serve to consolidate, focus and localize E2 and/or phosphoryl signaling among distinct subcellular domains. Acknowledgements  We appreciate very much the assistance of Tim Hewison (Grieg Seafood B.C. Ltd., Campbell River, B.C.) and Lance Page (Marine Harvest Canada, Duncan, B.C.) for their provision of the testis and sperm samples, respectively. We also would like to thank Dr. Duncan Liew (EMD Millipore Corp.) and Dr. Sybille Rex (Abcam Incorp.) for their help in the selection of commercial antibodies compatible with a salmon study. We are grateful to James Nagler (University of Idaho, Moscow, ID) and Martin Tresguerres (University of California, San Diego, CA), for generously providing antibodies to trout ERα and sAC, respectively. The contribution of ribosomes from Bruno Klaholz (CBI-IGBMC, Illkirch, France) was invaluable to our control experiments for Y10b labeling. We thank Patrick Nahirney (University of Victoria, Faculty of Medicine) for access to his Jeol TEM. Compliance with ethical standards  Conflict of interest  The authors declare that they have no conflict of interest. Ethical approval  All applicable international, national, and/or institutional guidelines for the care and use of animals were followed. The animals used in this study underwent normal commercial broodstock practices (sperm retrieval) or destined for food production (testis extraction). Funding  This research was supported by a Natural Resources and Applied Sciences Team Grant from the B.C. Innovation Council (WSD, BFK) and the Natural Sciences and Engineering Research Council of Canada (BFK, WSD).

References Aquila S, Sisci D, Gentile M, Middea E, Siciliano L, Andò S (2002) Human ejaculated spermatozoa contain active P450 aromatase. J Clin Endocrinol Metab 87:3385–3390 Aquila S, Sisci D, Gentile M, Middea E, Catalano S, Carpino A, Rago V, Andò S (2004) Estrogen receptor (ER)α and ERβ are both expressed in human ejaculated spermatozoa: evidence of their direct interaction with phosphatidylinositol-3-OH kinase/Akt pathway. J Clin Endocrinol Metab 89:1443–1451 Beltrán C, Vacquier VD, Moy G, Chen Y, Buck J, Levin LR, Darszon A (2007) Particulate and soluble adenylyl cyclases participate in

the sperm acrosome reaction. Biochem Biophys Res Commun 358:1128–1135 Cappallo-Obermann H, Schulze W, Jastrow H, Baukloh V, Spiess A-N (2011) Highly purified spermatozoal RNA obtained by a novel method indicates an unusual 28S/18S rRNA ratio and suggests impaired ribosome assembly. Mol Hum Reprod 17:669–678 Caulier M, Brion F, Chadili E, Turies C, Piccini B, Porcher J-M, Guiguen Y, Hinfray N (2015) Localization of steroidogenic enzymes and Foxl2a in the gonads of mature zebrafish (Danio rerio). Comp Biochem Physiol A Mol Integr Physiol 188:96–106 Chen J-Q, Yager JD, Russo J (2005) Regulation of mitochondrial respiratory chain structure and function by estrogens/estrogen receptors and potential physiological/pathophysiological implications. Biochim Biophys Acta 1746:1–17 Chennathukuzhi V, Morales CR, El-Alfy M, Hecht NB (2003) The kinesin KIF17b and RNA-binding protein TB-RBP transport specific cAMP-responsive element modulator-regulated mRNAs in male germ cells. Proc Natl Acad Sci USA 100:15566–15571 Delalande C, Goupil A-S, Lareyre J-J, Le Gac F (2015) Differential expression patterns of three aromatase genes and of four estrogen receptors genes in the testes of trout (Oncorhynchus mykiss). Mol Reprod Dev 82:694–708 Dı́ez-Sánchez C, Ruiz-Pesini E, Montoya J, Pérez-Martos A, Enrı́quez JA, López-Pérez MJ (2003) Mitochondria from ejaculated human spermatozoa do not synthesize proteins. FEBS Lett 553:205–208 Eddy EM, Toshimori K, O’Brien DA (2003) Fibrous sheath of mammalian spermatozoa. Microsc Res Tech 61:103–115 Fang P, Zeng P, Wang Z, Liu M, Xu W, Dai J, Zhao X, Zhang D, Liang D, Chen X, Shi S, Zhang M, Wang L, Qiao Z, Shi H (2014) Estimated diversity of messenger RNAs in each murine spermatozoa and their potential function during early zygotic development. Biol Reprod 90(5):94–111 Garden G, Canady K, Lurie D, Bothwell M, Rubel E (1994) A biphasic change in ribosomal conformation during transneuronal degeneration is altered by inhibition of mitochondrial, but not cytoplasmic protein synthesis. J Neurosci 14:1994–2008 Gohin M, Bodinier P, Fostier A, Chesnel F, Bobe J (2011) Aromatase is expressed and active in the rainbow trout oocyte during final oocyte maturation. Mol Reprod Dev 78:510–518 Guiguen Y, Fostier A, Piferrer F, Chang C-F (2010) Ovarian aromatase and estrogens: a pivotal role for gonadal sex differentiation and sex change in fish. Gen Comp Endocrinol 165:352–366 Gur Y, Breitbart H (2008) Protein synthesis in sperm: dialog between mitochondria and cytoplasm. Mol Cell Endocrinol 282:45–55 Hammes SR, Levin ER (2007) Extranuclear steroid receptors: nature and actions. Endocr Rev 28:726–741 Haramoto Y, Takahashi S, Asashima M (2006) Two distinct domains in pro-region of Nodal-related 3 are essential for BMP inhibition. Biochem Biophys Res Commun 346:470–478 Hess KC, Jones BH, Marquez B, Chen Y, Ord TS, Kamenetsky M, Miyamoto C, Zippin JH, Kopf GS, Suarez SS, Levin LR, Williams CJ, Buck J, Moss SB (2005) The “soluble” adenylyl cyclase in sperm mediates multiple signaling events required for fertilization. Dev Cell 9:249–259 Holdcraft RW, Braun RE (2004) Hormonal regulation of spermatogenesis. Int J Androl 27:335–342 Inaba K (2003) Molecular architecture of the sperm flagella: molecules for motility and signaling. Zool Sci 20:1043–1056 Ingraham HA, Hirokawa Y, Roberts LM, Mellon SH, McGee E, Nachtigal MW, Visser JA (2000) Autocrine and paracrine Müllerian inhibiting substance hormone signaling in reproduction. Recent Prog Horm Res 55:53–67 Johnson GD, Sendler E, Lalancette C, Hauser R, Diamond MP, Krawetz SA (2011) Cleavage of rRNA ensures translational cessation in sperm at fertilization. Mol Hum Reprod 17:721–726

13

Johnson GD, Mackie P, Jodar M, Moskovtsev S, Krawetz SA (2015) Chromatin and extracellular vesicle associated sperm RNAs. Nucleic Acids Res 43:6847–6859 Kikuchi K, Hamaguchi S (2013) Novel sex-determining genes in fish and sex chromosome evolution. Dev Dyn 242:339–353 Kimmins S, Kotaja N, Davidson I, Sassone-Corsi P (2004) Testisspecific transcription mechanisms promoting male germ-cell differentiation. Reproduction 128:5–12 Kleene KC (2013) Connecting cis-elements and trans-factors with mechanisms of developmental regulation of mRNA translation in meiotic and haploid mammalian spermatogenic cells. Reproduction 146:R1–R19 Kotaja N, Sassone-Corsi P (2007) The chromatoid body: a germ-cellspecific RNA-processing centre. Nat Rev Mol Cell Biol 8:85–90 Krawetz SA, Kruger A, Lalancette C, Tagett R, Anton E, Draghici S, Diamond MP (2011) A survey of small RNAs in human sperm. Hum Reprod 26:3401–3412 Krisfalusi M, Miki K, Magyar PL, O’Brien DA (2006) Multiple glycolytic enzymes are tightly bound to the fibrous sheath of mouse spermatozoa. Biol Reprod 75:270–278 Kumar RS, Trant JM (2001) Piscine glycoprotein hormone (gonadotropin and thyrotropin) receptors: a review of recent developments. Comp Biochem Physiol B Biochem Mol Biol 129:347–355 Liu Y, Dettin LE, Folmer J, Zirkin BR, Papadopoulos V (2007) Abnormal morphology of spermatozoa in cytochrome P450 17alphahydroxylase/17,20-lyase (CYP17) deficient mice. J Androl 28:453–460 Manandhar G, Schatten H, Sutovsky P (2005) Centrosome reduction during gametogenesis and its significance. Biol Reprod 72:2–13 Mao S, Goodrich RJ, Hauser R, Schrader SM, Chen Z, Krawetz SA (2013) Evaluation of the effectiveness of semen storage and sperm purification methods for spermatozoa transcript profiling. Syst Biol Reprod Med 59:287–295 McDonald KL (1994) Electron microscopy and EM immunocytochemistry. Methods Cell Biol 44:411–444 McIntosh CJ, Lun S, Lawrence S, Western AH, McNatty KP, Juengel JL (2008) The proregion of mouse BMP15 regulates the cooperative interactions of BMP15 and GDF9. Biol Reprod 79:889–896 Mizrahi R, Breitbart H (2014) Mitochondrial PKA mediates sperm motility. Biochim Biophys Acta 1840:3404–3412 Morisawa M, Ishida K (1987) Short-term changes in levels of cyclic AMP, adenylate cyclase, and phosphodiesterase during the initiation of sperm motility in rainbow trout. J Exp Zool 242:199–204 Morisawa S, Morisawa M (1988) Induction of potential for sperm motility by bicarbonate and pH in rainbow trout and chum salmon. J Exp Biol 136:13–22 Nagler JJ, Cavileer T, Sullivan J, Cyr DG, Rexroad C (2007) The complete nuclear estrogen receptor family in the rainbow trout: discovery of the novel ERα2 and both ERβ isoforms. Gene 392:164–173 Noda T, Shidara O, Harayama H (2012) Detection of the activator cAMP responsive element modulator (CREM) isoform ortholog proteins in porcine spermatids and sperm. Theriogenology 77:1360–1368 Pedram A, Razandi M, Deschenes RJ, Levin ER (2012) DHHC-7 and -21 are palmitoyl acyltransferases for sex steroid receptors. Mol Biol Cell 23:188–199 Pisarska MD, Kuo FT, Bentsi-Barnes IK, Khan S, Barlow GM (2010) LATS1 phosphorylates forkhead L2 and regulates its transcriptional activity. Am J Physiol Endocrinol Metab 299:E101–E109 Pisarska MD, Barlow G, Kuo F-T (2011) Minireview: roles of the forkhead transcription factor FOXL2 in granulosa cell biology and pathology. Endocrinology 152:1199–1208

13

Histochem Cell Biol Sarkar S, Jun S, Simpkins JW (2015) Estrogen amelioration of Aβ-induced defects in mitochondria is mediated by mitochondrial signaling pathway involving ERβ, AKAP and Drp1. Brain Res 1616:101–111 Schatten H, Sun Q-Y (2010) The role of centrosomes in fertilization, cell division and establishment of asymmetry during embryo development. Semin Cell Dev Biol 21:174–184 Schulz RW, de França LR, Lareyre J-J, LeGac F, Chiarini-Garcia H, Nobrega RH, Miura T (2010) Spermatogenesis in fish. Gen Comp Endocrinol 165:390–411 Sendler E, Johnson GD, Mao S, Goodrich RJ, Diamond MP, Hauser R, Krawetz SA (2013) Stability, delivery and functions of human sperm RNAs at fertilization. Nucleic Acids Res 41:4104–4117 Shiraishi E, Yoshinaga N, Miura T, Yokoi H, Wakamatsu Y, Abe S-I, Kitano T (2008) Müllerian inhibiting substance is required for germ cell proliferation during early gonadal differentiation in medaka (Oryzias latipes). Endocrinology 149:1813–1819 Skepper JN (2000) Immunocytochemical strategies for electron microscopy: choice or compromise. J Microsc 199:1–36 Solakidi S, Psarra AM, Nikolaropoulos S, Sekeris CE (2005) Estrogen receptors alpha and beta (ERalpha and ERbeta) and androgen receptor (AR) in human sperm: localization of ERbeta and AR in mitochondria of the midpiece. Hum Reprod 20:3481–3487 Tresguerres M, Barott KL, Barron ME, Roa JN (2014) Established and potential physiological roles of bicarbonate-sensing soluble adenylyl cyclase (sAC) in aquatic animals. J Exp Biol 217:663–672 Turner RM (2003) Tales from the tail: what do we really know about sperm motility? J Androl 24:790–803 Vadnais ML, Cao W, Aghajanian HK, Haig-Ladewig L, Lin AM, AlAlao O, Gerton GL (2014) Adenine nucleotide metabolism and a role for AMP in modulating flagellar waveforms in mouse sperm. Biol Reprod 90(6):128. doi:10.1095/biolreprod.113.114447 Valsecchi F, Ramos-Espiritu LS, Buck J, Levin LR, Manfredi G (2013) cAMP and mitochondria. Physiology 28:199–209 Viñas J, Piferrer F (2008) Stage-specific gene expression during fish spermatogenesis as determined by laser-capture microdissection and quantitative-PCR in sea bass (Dicentrarchus labrax) gonads. Biol Reprod 79:738–747 von Schalburg KR, Rise ML, Brown GD, Davidson WS, Koop BF (2005) A comprehensive survey of the genes involved in maturation and development of the rainbow trout ovary. Biol Reprod 72:687–699 von Schalburg KR, Yasuike M, Yazawa R, de Boer JG, Reid L, So S, Robb A, Rondeau EB, Phillips RB, Davidson WS, Koop BF (2011) Regulation and expression of sexual differentiation factors in embryonic and extragonadal tissues of Atlantic salmon. BMC Genomics 12:31–49 von Schalburg KR, Gowen BE, Rondeau EB, Johnson NW, Minkley DR, Leong JS, Davidson WS, Koop BF (2013) Sex-specific expression, synthesis and localization of aromatase regulators in one-year-old Atlantic salmon ovaries and testes. Comp Biochem Physiol B Biochem Mol Biol 164:236–246 von Schalburg KR, Gowen BE, Messmer AM, Davidson WS, Koop BF (2014) Sex-specific expression and localization of aromatase and its regulators during embryonic and larval development of Atlantic salmon. Comp Biochem Physiol B Biochem Mol Biol 168:33–44