ZEBRAFISH Volume 11, Number 2, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/zeb.2013.0938
Temporal Dynamics of Oocyte Growth and Vitellogenin Gene Expression in Zebrafish (Danio Rerio) Michelle H. Connolly,1 Rachel M. Dutkosky,1 Tze P. Heah,1 Gary S. Sayler,1–3 and Theodore B. Henry1,4,5
Little is known about how hepatic vitellogenin gene (vtg) expression relates to oogenesis in fish, especially among fractional spawners. The objective of this study was to relate hepatic vtg 1A/B expression to stagespecific oocyte development in zebrafish (Danio rerio), an asynchronous spawning fish. Liver samples were collected at seven time points postspawning (1–32 days) and fish were preserved for subsequent histological analyses. Relative vtg 1A/B expression among liver samples was quantified by reverse transcription– quantitative PCR and oogenesis was evaluated following standard hematoxylin and eosin staining of serial ovarian sections. Histological analyses indicate that a subset of previtellogenic oocytes (stages 1–2) transitioned into postvitellogenic oocytes (stages 3–4) within 4 days (96 h) postspawning. By 8 days postspawning (192 h), the majority of the ovary was occupied by mature (stage 4) oocytes, a trend that continued through 32 days postspawning. Hepatic vtg 1A/B gene expression was upregulated 3.89-fold 1-h postspawning relative to the average gene expression across all time points, but was not correlated to stage-specific oogenesis. Follicular atresia among fish sampled 32 days postspawning highlights the importance of regular spawning in zebrafish and suggests that the event of spawning itself may be integral to the regulation of oocyte development.
he production of yolky eggs involves a series of developmental events that ultimately transform oogonia into mature ova. It is during this process that egg yolk precursor proteins (vitellogenins) are deposited into developing oocytes. Among fish, the production of yolky eggs consists of four stages of development characterized by primary oocyte growth, cortical alveolus development, vitellogenesis, and maturation.1 The transition from one stage to the next may be seasonal (synchronous), punctuated (group synchronous), or continuous (asynchronous), depending on the species in question. Yet, all are coordinated by common reproductive hormones, which influence gene expression and mediate morphological change.2 In particular, pituitary gonadotropins control the delicate balance between the production of gonadal estrogens and hepatically derived vitellogenins, which together regulate oocyte growth and development.3 Vitellogenins are phospholipid glycoproteins that are of particular interest to fish biologists as they contribute to the production of yolk proteins that are vital to embryonic development.4 It should be noted that while vitellogenin genes
(vtgs) are strongly upregulated in the liver, they are also weakly expressed in other fish tissues, including the ovary, intestine, and white adipose tissue5,6; yet, little is known about what drives their expression. By contrast, hepatic vtg has been measured in many group synchronous spawners, including the Japanese eel (Anguilla japonica),7 rainbow trout (Oncorhynchus mykiss),8 plaice (Pleuronectes platessa),9 and goldfish (Carassius auratus).10 These studies have shown that synchronous spawning fish are characterized by well-defined vtg expression patterns, which closely parallel seasonal changes in steroid hormones.11–14 However, little is known about how vtg expression may vary among asynchronous spawning fish,15 especially in response to stage-specific oocyte development. Recent studies indicate that some fish have as many as three forms of vitellogenin (VgA, VgB, and VgC), which are coded by one or more vitellogenin gene(s). The multiplicity of vitellogenin is thought to have arisen due to whole genome duplication(s) during the evolution of the teleost lineage (*350 mya).16 In particular, zebrafish (Danio rerio), have been shown to express seven vtg genes (vtg 1–7), which code for VgA (vtg1, 4, 5, 6, 7), VgB (vtg2), and VgC (vtg3)
Center for Environmental Biotechnology, University of Tennessee, Knoxville, Tennessee. Departments of 2Microbiology, 3Ecology and Evolutionary Biology, and 4Forestry, Wildlife and Fisheries, University of Tennessee, Knoxville, Tennessee. 5 School of Life Sciences, Heriot-Watt University, Edinburgh, United Kingdom.
proteins.17 VgA and VgB have also been shown to contain a complete yolk protein domain, while VgC has a greatly reduced or missing polyserine phosvitin (Pv) protein domain.16 Notably, vtg genes coding for VgA and VgB have been shown to be responsive to exogenous estrogens6 and have thus been recognized as biomarkers of estrogen exposure in fish.16,18,19 Limited sequence divergence among the protein subtypes has made it difficult to distinguish VgA from VgB in many fish16,20,21; therefore, previous research efforts have focused on developing primer/probe sets that investigate genes that code for both VgA and VgB subtypes (e.g., vtg1, 4, 5, 6, 7; hereafter vtg 1A/B).22 The purpose of this study was to evaluate patterns of oocyte development and vtg 1A/B expression in D. rerio to further decipher the regulation of asynchronous ovarian development in this vertebrate model. Stage-specific oogenesis was assessed among histological sections of fish sampled 1–32 days postspawning, while hepatic vtg 1A/B expression was evaluated by real-time reverse transcription-PCR (RT-PCR). Gaining a greater understanding of the molecular and morphological patterns that underlie asynchronous ovarian development will contribute to the interpretation of differential gene expression among fish that share this reproductive strategy.
CONNOLLY ET AL.
in Bouin’s fixative for 48 h as per standard methods.29 Samples were subsequently transferred to 70% EtOH and prepared for histological analyses as described below. Histology
Transverse histological sections of the trunk were prepared to enable simultaneous examination of the paired ovaries. To capture the same region among all fish, the head and tail were removed postfixation by cutting away tissues posterior to the operculum and anterior to the dorsal fin (Fig. 1B). The remaining trunk tissue was subsequently bisected to expose the midregion of the ovary. Samples were decalcified for 2 h in Cal-Ex (Fisher) and rinsed several times in 70% EtOH according to standard procedures29 before embedding in paraffin wax. Similarly, two additional female fish were embedded sagitally to permit evaluation of the distribution of oocyte stages and ovary morphology from an anterior to posterior aspect (Fig. 2). Specimens were sectioned (5–6 lm) with an American Optical (Leica) microtome, mounted on glass slides, and stained with hematoxylin and eosin (H&E) according to standard procedures.29 Sections were examined under light microscopy using a Zeiss Stemi SV11 stereomicroscope and imaged with an Axiocam HRc Camera.
Materials and Methods Experimental fish
D. rerio were obtained from the Zebrafish Research Facility in the Center for Environmental Biotechnology at the University of Tennessee (UT), where fish were maintained according to standard conditions.23 Briefly, D. rerio were maintained at 28C – 0.5C under a 14D:10L cycle and fed a mixture of Artemia nauplii (Brine Shrimp Direct) and Tetramin flake food twice daily. All procedures were conducted with approval from the UT Institutional Animal Care and Use Committee.
Four stages of oocyte development were identified among imaged sections and color coded using Adobe Photoshop Elements 5.0 (Fig. 1). Oocytes were staged according to Selman et al., (1993),30 where stage 1 were small (7– 140 lm), basophilic, and had relatively large nuclei; stage 2 (140–340 lm) were characterized by the presence of cortical alveoli, a tripartite zona radiata (vitelline envelope), and a central germinal vesicle; stage 3 (340–690 lm) contained internal vitellus droplets and peripheral yolk vesicles; and stage 4 were large (0.69–0.75 lm) and filled with noncrystalline yolky bodies (Fig. 1C). As described in Selman et al.,30 the distinction between stages 1A/1B and 4/5 is only possible under electron microscopy. Therefore, for the purpose of the present study, the primary growth stage was considered stage 1, while newly ovulated oocytes were considered stage 4 (mature oocytes). ¨ c¸u¨ncu¨ and Atretic follicles were identified as described by U 31 Cxakıcı and defined as stage 5 in the present study. Once all oocytes were categorized according to stage, the number, proportion, and percent cover of each stage were computed for each fish using Nikon Elements Software (NIS). In accordance with previous organ-specific gene expression studies,32–34 we sampled three fish per time point (N = 21) to minimize the lethal sampling associated with harvesting hepatic tissue.
D. rerio (age *180 days) were successfully pair spawned for several weeks prior to the onset of the current study. At the end of a photoperiod, a single male and female fish were placed in a breeding chamber and separated by a partition. The partition was lifted at the onset of the next photoperiod and fish were allowed to spawn for 30 min. Time zero was determined by the onset of the spawning event. Twenty-one female D. rerio [mean total length 32.56 mm – 2.47SD; and weight 0.3216 g – 0.09SD], which had produced an average of 75 or more viable embryos per spawn, were selected and maintained in the absence of male fish. Three female D. rerio were randomly sampled at one of seven time points (1 h, 1 day, 2 days, 4 days, 8 days, 16 days, and 32 days) postspawning, euthanized in ice water, and dissected as described by Gupta and Mullins.24 Whereas wild zebrafish have been known to spawn every 2–3 days,25 we selected the above stated sampling design to characterize the 7–10 days breeding schedule that is common to many research settings,26–28 as well as examine the molecular and morphological repercussions of an extended spawning schedule. A small liver sample was subsequently excised, flash frozen in liquid nitrogen, and stored at - 80C until further processing. Following excision, the entire fish was immersed
RNA extraction and RT-quantitative PCR
RNA extraction, RT of RNA to cDNA, and quantitative PCR (qPCR) analyses were conducted as per Henry et al.22 with a few modifications as described below. Liver samples were briefly sonicated in the RNeasy (RLT) lysis buffer and filtered through QIAshredder (QIAGEN) spin columns. RNA extraction was conducted using a QIAGEN RNAeasy mini kit that included a DNase step to remove genomic
OOCYTE GROWTH AND VITELLOGENIN GENE EXPRESSION
FIG. 1. Transverse section of a female zebrafish 16 days postspawning (A–D) double stained with hematoxylin and eosin (HE). Zebrafish ovaries (Ov) were isolated and embedded in situ (B). Head and tail regions were removed by cutting away tissues posterior to the operculum and anterior to the dorsal fin [dotted lines, (B)]. The remaining trunk tissue was cut in half and sectioned transversely to capture the midregion of the ovary as shown in (A). Oocyte stages (1–4) in (C) have been color coded in (D), where stage 1 are shown red, stage 2 in yellow, stage 3 in blue, and stage 4 in green. Stage 5 (atresia, orange) can occur at any point during oogenesis, but the purpose of this study, has been categorized as stage 5 (see text for details). SB, swim bladder. Scale bar in (A), 1.0 mm, 250 lm in (C) and (D). Color images available online at www.liebertpub.com/zeb
DNA. Samples were subsequently eluted in 40 lL of nuclease-free water (Ambion) and assessed for sample purity (260/280 absorbance ratio), integrity (260/230 absorbance ratio), and concentration (ng/lL) using a NanoDrop ND1000 spectrophotometer (Themo Scientific) before storage at - 80C. Five hundred nanograms of purified RNA ( > 1.9 260/230 absorbance ratio) was subsequently reversed transcribed using a superscript III first strand synthesis kit (Invitrogen). Vtg transcripts were evaluated using previously published TaqMan primer/probe sets22 designed to assess the expression of vtg 1A/B (vtg1, 4, 5, 6, and 7, coding for VgA and VgB) and b-actin. Samples were diluted (1/32) and added to a Quantitect qRT PCR mixture (QIAGEN) containing 300 nM forward and reverse primers, 100 nM FAM-labeled probe, nuclease-free water, and master mix. Transcripts were amplified using a Chromo 4 Thermal Cycler (MJ Research) under the following conditions: 50C for 2 min, 95C for
15 min, followed by 45 cycles of 95C for 15 s and 60C for 1 min. PCR efficiencies were calculated from the slope of a dilution curve (E = [(10(slope/ - 1))–1]) generated for each gene on each plate and ranged from 84.2% to 85.1% (R2 = 98.9– 99.1). Finally, fold change was calculated relative to the average gene expression across all time points using the standard curve method ([Evtg 1A/BDCt vtg 1A/B]/[E b-actinDCt b-actin]) as described by Pfaffl.35 Statistical analysis
Data are presented as mean – 1 standard deviation unless otherwise noted. Following an initial assessment of normality (Shapiro–Wilk), data were analyzed using nonparametric statistics, as normality among histological samples could not be corrected by transformation. Wilcoxon’s Signed-Rank Test was used to assess vtg 1A/B expression and stage-specific
FIG. 2. Sagittal section of a female zebrafish (A) showing the relative location and homogeneous distribution of oocyte stages throughout the ovary (B). Scale Bar 1.0 mm. Color images available online at www.liebertpub.com/zeb
histological measurements (% oocyte number, % oocyte cover) relative to time postspawning. Statistical differences between paired means were measured using Student t-tests. General linear regression and Spearman’s Rank Correlation Coefficient (Spearman’s Rho) were also used to assess the relationship within and between stages of oogenesis as well as between vtg 1A/B gene expression and oogenesis 1–32 days postspawning. All analyses were computed using JMP (9.0.2) statistical software, where significance was assessed at p < 0.05. Results Fish reproduction and histology
Females produced an average of 164 eggs – 52SD per spawn, with a mean hatching rate of 89.2% (148 eggs –
FIG. 3. Change in ovary size and composition over time. Mean area ( – 1 SE) occupied by oocytes 1–32 days postspawning (A). Outlier sampled at 32 days [open circle, (A)], characterized by severely regressed ovaries (see also Fig. 5A). Mean ( – 1 SE) fold change in vtg 1A/B gene expression 1–32 days postspawning [gray bars in (B)] relative to the proportion of stages 1–2 (open squares) and 3–4 oocyte (closed squares). Percent cover of oocyte stages 1–2 [black circles, (C)], 3–4 [gray circles, (C)], and atretic follicles (open circles) 1–32 days postspawning, with 1 SE shown around each mean.
CONNOLLY ET AL.
52SD per clutch). Histological analyses of sagittal sections indicate that D. rerio ovaries contained a mixture of all stages of oogenesis throughout their length (Fig. 2). Therefore, subsequent analyses were conducted on transverse sections to capture both the left and right lobes of the D. rerio ovary. The mean cross-sectional area occupied by all stages of oogenesis (an approximation of ovary size) increased 2.89fold over the course of this study (1–32 days postspawning; Fig 3A). Notably, variations in ovary size increased 8–32 days [192–768 h] postspawning in association with the development of stage 4 oocytes. A period of developmental stasis preceded the onset of vitellogenesis, where the relative number of stage 1–2 oocytes remained stable through 4 days postspawning before developing into stage 3–4 oocytes (Fig. 3B). In addition, the transition between previtellogenic (stage 1–2) and vitellogenic (stage 3–4) oocytes was marked by an increase in atretic (stage 5) follicles (open circles, Fig. 3C). By 8 days postspawning (192 h), 69.0% – 9.9%SD of the ovary was occupied by mature oocytes, a trend that continued through 32 days postspawning, when 82.7% – 10.0%SD of the cross-sectional area contained stage 4 oocytes. Histological analyses of transverse sections indicate that most fish (20/21) contained all stages of oogenesis 1–32 days postspawning (but see exception below). The mean number of oocytes per section was 136 – 45SD, which consisted primarily of stage 1 (70.8%) and 2 (15.8%) oocytes. Most ovaries contained relatively few stage 3 oocytes (