Physiol. Genomics AJP: Advances in Physiology Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Physiol. Genomics 30: 111-122, 2007. First published March 20, 2007; doi:10.1152/physiolgenomics.00284.2006
1094-8341/07 $8.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Tables
Right arrow All Versions of this Article:
30/2/111    most recent
00284.2006v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Santos, E. M.
Right arrow Articles by Tyler, C. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Santos, E. M.
Right arrow Articles by Tyler, C. R.
Received 21 December 2006; accepted in final form 18 March 2007.
Physiological Genomics 30:111-122 (2007)
1094-8341/07 $8.00 © 2007 American Physiological Society

Molecular basis of sex and reproductive status in breeding zebrafish

E. M. Santos1, V. L. Workman2, G. C. Paull1, A. L. Filby1, K. J. W. Van Look3, P. Kille2,* and C. R. Tyler1,*

1 School of Biosciences, University of Exeter, Exeter
2 Cardiff School of Biosciences, Cardiff
3 Institute of Zoology, Zoological Society of London, London, United Kingdom


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 GRANTS
 REFERENCES
 
The zebrafish (Danio rerio) is used extensively as a model species for studies on vertebrate development and for assessing chemical effects on reproduction. Despite this, the molecular mechanisms controlling zebrafish reproduction are poorly understood. We analyzed the transcriptomic profiles of the gonads of individual zebrafish, using a 17k oligonucleotide microarray, to define the molecular basis of sex and reproductive status in sexually mature fish. The gonadal transcriptome differed substantially between sexes. Among the genes overexpressed in females, 11 biological processes were overrepresented including mitochondrion organization and biogenesis, and cell growth and/or maintenance. Among the genes overexpressed in males, six biological processes were overrepresented including protein biosynthesis and protein metabolism. Analysis of the expression of gene families known to be involved in reproduction identified a number of genes differentially expressed between ovaries and testes including a number of sox genes and genes belonging to the insulin-like growth factor and the activin-inhibin pathways. Real-time quantitative PCR confirmed the expression profiles for nine of the most differentially expressed genes and indicated that many transcripts are likely to be switched off in one of the sexes in the gonads of adult fish. Significant differences were seen between the gonad transcriptomes of individual reproductively active females reflecting their stage of maturation, whereas the testis transcriptomes were remarkably similar between individuals. In summary, we have identified molecular processes associated with (gonadal) sex specificity in breeding zebrafish and established a strong relationship between individual ovarian transcriptomes and reproductive status in females.

Danio rerio; gonad; reproduction; individual transcriptome; phenotypic anchoring


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 GRANTS
 REFERENCES
 
IN VERTEBRATES, REPRODUCTION is controlled by the brain-pituitary-gonadal (BPG) axis and involves a complex cascade of hormones and peptides with elaborate feedback mechanisms and with links to other biochemical pathways that control growth and metabolism (3). However, our understanding of the molecular pathways underlying reproductive development, even in mammals, is limited.

The zebrafish, Danio rerio, is used extensively as a model species for studies on development, reproduction, disease, and ecotoxicology. Its reproductive biology is similar to many other cyprinid species (one of the largest families of fish inhabiting freshwaters in the northern hemisphere), and this, together with its ease of culture in the laboratory, its short generation time (3–4 mo), and, crucially, the available genomic resources, makes it one of the most appropriate models to develop our understanding of the mechanisms controlling reproduction in fish. In recent years, the number of putative genes involved with reproductive development in fish and other animals has increased considerably. The generation of tissue-specific gene libraries for zebrafish, including for ovary and testis, have facilitated investigations into sex assignment and sexual development in this species (20, 23, 55). Extensive lists of transcripts have been generated in zebrafish that are overexpressed in adult females when compared with gene expression profiles in adult males (49). Many of the genes identified have corresponded to well-known female-specific genes, but a large proportion have been either novel or unidentified. These studies, however, were performed using RNA derived from whole body preparations and pooled from a number of fish, thus no relationship was established between gene expression and the physiology of the individuals (49). Comparisons of the list of female enriched genes with the equivalent list generated in Drosophila have found that the germ line, but not the somatic sex-enriched genes, is conserved between vertebrate and invertebrate species, indicating that the process of gametogenesis is highly conserved during evolution (49).

In mammals, investigations into sex-specific gene expression profiles in gonads have been more extensive. Employing a mouse 30,000 gene microarray, Rinn and colleagues (38) identified >500 genes upregulated in testes and 300 genes upregulated in ovaries. Immune-response genes were underrepresented in the testis, but the functional significance of this was not determined. In ovary, the only ontological category that was significantly overrepresented was the monooxygenase enzymes, which are part of the P450 family involved in steroid hormone biosynthesis and steroid and drug metabolism. Interestingly, sexual dimorphism in gene expression was also found for the kidney, where the expression of 27 genes was found to differ between sexes (38). Similarly, studies in rat have shown sex-specific differences in gene expression in the liver. The differences observed were reported to be driven by differences between the sexes in the pattern of pituitary secretion of growth hormone (1), reviewed in Ref. 39.

In this study, we investigated the molecular basis for the phenotypic specificity between sexes and between individuals within a sex in breeding zebrafish. To do so, we searched for sex-related differences in the transcriptome of reproductively active zebrafish at the level of the gonads. We further analyzed for associations between molecular profiles and reproductive status between individuals for both males and females.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 GRANTS
 REFERENCES
 
Chemicals
All chemicals and reagents were obtained from Sigma-Aldrich, Poole, UK, unless stated otherwise.

Array Fabrication
Preparation of spotted glass oligonucleotide microarrays.
Oligonucleotide microarrays were produced at the Cardiff University Molecular Biology Support Unit (Cardiff, UK). Oligonucleotides were purchased from the Sigma-Genosys (Cambridge, UK) Zebrafish Oligolibrary (65-mer with 5'-C6 amino modifier, 500 pmol in Genetix X7020) and printed on Amersham CodeLink slides (Amersham Biosciences, Little Chalfont, UK) at a final concentration of 25 µM in 50 mM sodium phosphate, pH 8.5 buffer. In addition to the standard library, 167 custom-designed oligos were included to represent genes of known involvement in the control of reproductive development. Custom oligos were designed based principally on the conserved regions of the target genes in other fish species or other vertebrate species in the cases where insufficient sequence data were available in fish. The Lucidea ScoreCard (Amersham) was spotted 10 times on each slide, and the oligo library was spotted once using a SpotArray72 (Perkin Elmer, Boston, MA), supplied with 48 SMP3 Microspotting pins (Telechem, Sunnyvale, CA). After printing, the microarrays were placed in a saturated NaCl chamber (humidity ~75%) overnight at room temperature and then stored in the dark and under a vacuum, until required.

Prehybridization of oligonucleotide glass arrays.
Immediately prior to use, the slides were blocked in 1% BSA dissolved in 5x SSC and 0.1% SDS for 45 min at 42°C. The blocking buffer was removed by washing 10 times in water followed by two times in isopropanol, and the slides were then air dried to remove all traces of alcohol prior to the hybridization.

Biological Analysis
Experimental design.
Gonadal gene expression profiles were analyzed (using an oligonucleotide microarray) and related to sex and reproductive status of the individuals (determined via histological analysis of the gonads) in a breeding colony of zebrafish. To do so, microarrays were conducted on amplified RNA extracted from the gonads of individual fish. The microarray experiments followed a reference design in which each sample was hybridized in competition with a pooled sample of RNA derived from all individuals in the study. Gene expression profiles were compared between sexes (males and females) and also between individuals within each sex (stages of reproductive development).

Animals.
A colony of adult breeding zebrafish (6 males and 6 females) was kept in a 15-l tank under flow-through conditions (10 l/day) under a 12 h light-dark cycle with an artificial dawn-dusk transition of 30 min, at 28 ± 1°C. Fish were fed to satiation twice daily, each morning on freshly hatched Artemia naupli, and each afternoon on Tetramin dry tropical flake food (Tetramin; Tetrawerke, Melle, Germany). The number of eggs produced and embryo survival to 8 h postfertilization were monitored daily over a period of 40 days to determine the breeding performance of the colony. At the end of this monitoring period, the fish were killed humanely by an overdose of benzocaine (according to UK Home Office regulations), and the wet weights and fork lengths of the individual fish were recorded. One of the gonads from each fish was extracted and stored in RNAlater (Ambion, Huntingdon, UK) at –20°C, and the second gonad was extracted and processed for histological analysis.

Histological analysis.
Individual gonads were fixed in Bouin's solution (Raymond A. Lamb, Eastbourne, UK) for 4 h and subsequently washed twice in 70% industrial methylated spirit (IMS). The samples were dehydrated in IMS up to 100% and embedded in paraffin wax (Merck Eurolab, Poole, UK), using a Shandon tissue processor (Citadel 2000). Serial sections were cut at 5 µm, floated in a water bath, and collected onto glass slides. Sections were stained with Harris's hematoxylin and eosin (Merck Eurolab), treated with DPX mountant (Merck Eurolab), and analyzed by light microscopy (Zeiss Axioskop 40 microscope; Carl Zeiss, Oberkochen, Germany), and digital images were obtained using an Olympus DP70 charge-coupled device camera (Olympus Optical) coupled to analySIS 3.2 software (Soft Imaging System, Munster, Germany). A minimum of six gonad sections were analyzed for each fish. Testes were analyzed for the presence of the different stages of spermatogenesis (spermatogonia, spermatocytes, spermatides, and spermatozoa) and categorized by the most advanced stage of sex cell development. Ovaries were analyzed for the presence and relative abundance of the different stages of oocyte development (primary oocytes, cortical alveolar stage, previtellogenic oocytes, vitellogenic oocytes, and postovulatory follicles).

RNA extraction.
RNA was extracted from the gonad samples with TRI reagent, according to the manufacturer's instructions. In addition to the isopropanol precipitation step, a lithium chloride precipitation was performed. The RNA was then resuspended in water, and the quality was evaluated by gel electrophoresis, on a 0.7% agarose gel.

RNA amplification and labeling.
Due to the limited amount of RNA obtained from individual ovaries and testes, an amplification procedure was employed prior to hybridization onto the microarrays. Total RNA (0.1–2 µg) was amplified using the Amino Allyl MessageAmp aRNA amplification kit (Ambion), following the manufacturer's instructions. Control RNA spikes provided with the Lucidea ScoreCard were amplified simultaneously to evaluate the possible bias introduced by the amplification procedures. The quality of the resulting amino allyl-labeled antisense RNA (aaRNA) was verified by gel electrophoresis and by measuring the absorbance in a GeneQuant Spectrophotometer (Amersham). For both ovaries and testes, 5 µg of each sample were pooled together to generate a reference sample. aaRNA (5 µg) of the reference and each experimental sample was labeled with Cy5 and Cy3, respectively, using the FluoroLink Cy5 and Cy3 monofunctional dye (Amersham), according to the instructions provided within the Amino Allyl MessageAmp aRNA amplification kit.

Quality control of the labeled target.
Two methods were employed to verify the quality of the labeled aaRNA before hybridization. In the first method, the absorbance of the labeled aaRNA was measured at 260 nm, to quantify the total amount of RNA and at 550 and 650 nm to measure the amount of incorporated Cy3 and Cy5 label, respectively (using an Ultrospec 2100 pro UV-VIS spectrophotometer, Amersham). The following equation was employed to determine the percentage of incorporation (amount of incorporated dye x 350/amount of RNA), and samples with values >70 or <20 were rejected. The second method for assessing RNA quality employed gel electrophoresis (1.5% agarose gel in Tris-acetate-EDTA buffer). Gel trays were specially designed to the dimensions of a microscope slide to allow for its analysis with a microarray scanner. We diluted 1 µl of labeled aaRNA sample 1:2 with 50% glycerol, loaded it onto the gel, and ran it at 120 V for 20 min. Gels were scanned in a LSIV scanner (Genomic Solutions, Huntingdon, UK) at 550 and 675 nm for Cy3 and Cy5, respectively. The relative abundance of the incorporated and nonincorporated label, as well as the relative abundance of the short and long transcripts, was determined visually, and samples containing large amounts of nonincorporated label or high proportions of very short transcripts were considered of low quality and rejected for array hybridization.

Sample preparation for hybridization.
To optimize the ability of the labeled aaRNA to hybridize to the oligo probes spotted onto the array, samples were sonicated for 5 min (Kerry Ultrasonics, Skipton, UK). The size of the resulting transcripts was confirmed to be <600 bp by gel electrophoresis (see above). The labeled aaRNA samples were then evaporated to <10 µl in a Concentrator 5301 (Eppendorf, Cambridge, UK), and the final volume was adjusted to 10 µl with water.

Hybridization.
Reference (10 µl) and sample (10 µl) were labeled with Cy5 and Cy3, respectively, and combined to make a total volume of 20 µl. The mixture was then denatured for 3 min at 95°C and rapidly cooled by centrifuging at 14,000 rpm for 1 min. We then added 20 µl of 2x hybridization buffer (50% formamide, 10x SSC, 0.2% SDS) to each sample and mixed that by gently pipetting. This was subsequently added to the blocked arrayed slide, and another slide (blank) was overlaid on the hybridization mixture. The slides were hybridized in a humidified airtight box for 36 h at 42°C.

Array washing.
After hybridization was complete, slides were separated by gently shaking in 4x SSC, following by washing for 5 min at 42°C in 2x SSC, 0.1% SDS. Slides were then transferred to 0.2x SSC and incubated for 1 min at room temperature, followed by a final wash in 0.1x SSC for 1 min. Slides were then air-dried and stored in an airtight container in the dark until scanned.

Scanning, feature extraction, and analysis.
Slides were scanned in a ScanArray Express HT scanner (Perkin Elmer) at 550 and 675 nm for Cy3 and Cy5, respectively, at a resolution of 10 µm. The photomultiplier tube voltage was set at 100% throughout for both the red and green channels. Feature extraction was done with Imagene 5 (BioDiscovery, El Segundo, CA) software. Spots with high local background or aberrant spot shape were flagged by the software and checked manually.

Real-time quantitative PCR.
Four genes overexpressed in testes [tubulin alpha 7 (tuba7), septin 4 (sept4), heat shock protein 70 (hsp70), and cyclin g2 (ccng2)], five genes overexpressed in ovaries [RNA binding protein with multiple splicing 2 (rbpms2), transmembrane phosphatase with tensin homology (tpte), sox11b, cyclin b2 (ccnb2), and connexin 44.2 (cx44.2)] in the array dataset, and amh, a gene involved in male sex differentiation, were selected for real-time PCR analyses. Primers specific for target genes were designed with Beacon Designer 3.0 software (Premier Biosoft International, Palo Alto, CA) and purchased from MWG-Biotech (Ebersburg, Germany). Primer pairs were optimized and validated for real-time quantitative PCR (RT-QPCR) as described in Ref. 11. Specificity of primer sets throughout this range of detection was confirmed by the observation of single amplification products of the expected size and Tm and sequence. All assays were quantitative, with standard curve [mean threshold cycle (Ct) vs. log cDNA dilution] slopes of between –2.693 and –3.464, translating to high efficiencies [E; E = 10(–1/slope) (36)] of 1.94–2.31. Over the detection range, the linear correlation (R2) between the mean Ct and the logarithm of the cDNA dilution was >0.98 in each case. Primer sequences, National Center for Biotechnology Information (NCBI) GenBank accession numbers, PCR product sizes, and annealing temperatures are shown in Table 1.


View this table:
[in this window]
[in a new window]

  		Table 1. Primers used for the RT-QPCR analyses

 
To determine the expression of the target mRNAs, cDNA was synthesized from 1 µg of RQ1 DNase-treated (Promega, Southampton, UK) total RNA using random hexamers (MWG-Biotech) and MMLV reverse transcriptase (Promega), following manufacturer's instructions. cDNA was diluted (1:3), and RT-QPCR using SYBR Green chemistry was performed with the iCycler iQ Real-time Detection System (Bio-Rad Laboratories, Hercules, CA) as described previously (11) with the appropriate annealing temperatures (Table 1). A template-minus negative control was run in triplicate on each plate, and a reverse transcriptase-minus negative control was carried out for each sample to verify the absence of cDNA or genomic DNA contamination, respectively. Efficiency-corrected relative expression levels were determined as in Ref. 11 by normalizing to the "housekeeping" gene ribosomal protein l8 (rpl8), which was measured in each sample. This gene was chosen as an housekeeping gene based on its established use as an internal control for RT-QPCR experiments in fish (10, 56). Furthermore, our array dataset confirmed that its expression did not change significantly between ovaries and testes. Intra-assay coefficient of variation (CV) was 2.42% (n = 96). Interassay CVs were not measured because all of the samples for each gene were run on the same plate.

Bioinformatic analysis.
Reporter annotation was performed by exploiting MegaBlast algorithm to identify the best match to the oligonucleotide sequences with cDNAs associated with release Zv5 version of the zebrafish genome (Sanger Center release July 2005), the Refseq cDNA dataset (NCBI release August 2005) and all deposited mRNAs/expressed sequence tags (EST) present in GenBank (August 2005) allowing a maximum of a 2-bp mismatch. GenBank entries were linked to their respective Unigene clusters (September 2005). Physical location was assigned from genome release Zv5 associated to the matching Ensembl gene identifier. Zfin IDs were linked to the reporters using the Ensembl gene identifier, or alternatively the NCBI Refseq identifier, if an Ensembl match had not been identified. Ontological annotation associated with the reporters took advantage of the European Bioinformatics Institute GOA Xref table, allowing linkage of Ensembl, GenBank, and Unigene identifiers to the relevant Uniprot Identifier that, in turn, can be linked to the associated Gene Ontology (GO) annotation term.

Prior to analyzing the data we preprocessed them in R using the LimmaGUI package, which performed background subtraction, followed by a pin-based Lowess within-array normalization and a scaled between chip normalization (50). The array data were imported into GeneSpring 7.0 software (Silicon Genetics) for further analysis. For each experimental group, data were filtered on "flags" and on intensity of the raw data, and each spot was required to be present and >0.01 in at least 40% of the conditions to be considered for further analysis. Principal component analysis (PCA) was performed to identify the main trends in gene expression in the gonadal datasets. Prior to the statistical analysis, data that changed <1.2-fold between conditions were also removed. To identify differentially expressed genes, comparisons between treatment groups were performed by t-test (P < 0.05) followed by multiple testing correction (Benjamini and Hochberg false discovery rate). In addition, the data were filtered on "fold change" to identify genes that were up- and downregulated between conditions. Gene lists were generated and clustered with both gene and condition trees using distance as similarity measure. The distribution of GO terms in the gene lists overexpressed in males and females were compared with the distribution of GO terms in the whole array to identify the biological processes overrepresented among the genes of interest. The full microarray dataset was deposited in the NCBI Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) with the series record GSE6063.

Throughout this paper, data are presented as means ± SE.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
GRANTS
 REFERENCES
 
Advances in the zebrafish genome sequencing project, and the interest in the reproductive biology of the zebrafish in the contexts of both fundamental and applied research, make it timely for the application of postgenomic technologies to develop a more comprehensive understanding of the reproductive physiology of this model organism. In this pursuit, here we employed a 17k oligonucleotide microarray and were able to identify sex-specific patterns of gene expression at the level of the gonad and to establish links between gene expression profiles and the reproductive biology of individuals in a breeding colony.

Biological Analysis
The fish used in the study had an average length and weight of 33.33 ± 0.33 mm and 0.44 ± 0.01 g for the males (n = 6) and 32.5 ± 0.85 mm and 0.47 ± 0.05 g for the females (n = 6), and this was consistent for sexually mature zebrafish of this strain (WIK) bred in our laboratory. During the 40 days prior to sampling, the spawning events followed a regular periodicity (with peaks in egg production every 2–4 days), and the average number of eggs spawned per female per day was 17.7 ± 2.0. The mean number of live embryos per female per day was 11.5 ± 1.5. Histological analysis showed that the proportion of follicles at the various stages of oogenesis varied markedly between individuals; in some females primary oocytes predominated (e.g., fish no. 12), and, at the other extreme, the ovaries of other females contained principally vitellogenic oocytes (e.g., fish nos. 9 and 10; Fig. 1). This observation is consistent with individual females within a breeding colony spawning every 2–3 days, but not necessarily in synchrony with other females in the colony (31). The presence of vitellogenic oocytes in all females, and of postovulatory follicles in all but one female, demonstrated that all females were actively reproducing. Similar histological analysis of the testes showed that all males were producing spermatozoa and contained germ cells at all stages of development (Fig. 1). It should be realized, however, that the reproductive capacity of the reproductively active males was not necessarily equal. Indeed, recent studies have indicated that significant differences can occur in sperm quality (motility) between individual males in a breeding colony (46). Furthermore, dominance hierarchies can exist and thus the relative contribution of males to the next generation can differ markedly (43). These features highlight the need for analysis of gene expression profiles of individuals, rather than on pooled samples, for establishing the molecular mechanisms affecting reproductive capacity. A pooled sample approach only, with its intrinsic constraints, has been adopted previously for studies on zebrafish (22, 24, 26).


Figure 1
View larger version (191K):
[in this window]
[in a new window]

  		Fig. 1. Histopathology of the gonads. A: testis lobule of a male containing all stages of spermatogenesis including spermatozoa (sg, spermatogonia; sc, spermatocytes; st, spermatids; sz, spermatozoa). Scale bar represents 50 µm. Ovary containing oocytes predominantly at early stages of development (B), including oocytes at advanced stages of development (C), containing large numbers of postovulatory follicles (D). po, Primary oocytes; vo, vitellogenic oocytes; pof, postovulatory follicle. Scale bar represents 200 µm.

 
Transcriptomic Profiles
Microarray analysis of individual gonads revealed that 8,769 genes were consistently expressed in the gonads from which 7,976 genes were consistently expressed in the ovaries and 7,060 genes were consistently expressed in the testes. This corresponded to between 42 and 53% of all probes represented in the microarray. This compares well with other published reports employing the same oligonucleotide set and same platform [e.g., ~35% of all genes were reported to be consistently expressed in the muscle in the studies conducted by Malek and colleagues (24)]. Analysis of a subset of the genes differentially expressed between the ovary and testis against the expression of the corresponding gene clusters available on Unigene Expression Profiles found that the trends were common in 81% of the cases. The genes chosen (genes differentially expressed between ovaries and testes; P < 0.001) included both genes with low expression in one or both tissues and genes for which the difference in expression between ovaries and testes was less than twofold. In both cases, the accuracy of the Unigene Expression Profiles might be expected to be relatively low because the number of ESTs submitted for that gene in the tissues of interest was low and often zero, or the number of ESTs submitted would not be sufficient to accurately distinguish between relatively small differences in gene expression. Despite this, the comparability of the microarray and Unigene datasets was very high, demonstrating the accuracy of our array measurements. The microarray data were further compared with RT-QPCR data for nine differentially expressed transcripts, and the two datasets followed the same trend for all nine genes (Fig. 5). The RT-QPCR data analysis was performed on the same individual RNA samples that were used for amplification and microarray hybridizations, but prior to the amplification. The fold difference in gene expression detected by RT-QPCR was greater than the fold difference in the microarray dataset in all cases (ranging between 2- and 456-fold). This may be due to the fact that for the highly expressed transcripts in the array, spot saturation can occur, leading to underestimation of the difference in gene expression between conditions in the array dataset. This fact does not detract from the biological conclusions drawn from the array dataset, however, because these transcripts would still appear as strongly differentially expressed between conditions. Due to the limited amount of RNA extracted from some of the biological samples analyzed (individual testis), an amplification procedure was required to generate enough targets for hybridization to the arrays. We employed a linear amplification method (in vitro transcription of double stranded cDNA), and, therefore, bias of the amplification towards the most abundant transcripts was unlikely. If bias were to occur then greater differences in expression of the differentially expressed genes would be expected when the same (original) RNA was analyzed by microarray compared with RT-QPCR. This was not the case. On the contrary, and for all transcripts analyzed, the fold-change in gene expression measured in the microarrays was consistently lower than the fold-change determined by RT-QPCR. Together, the comparisons between the gene expression profiles derived from the array datasets and both the Unigene Expression Profiles and gene expression obtained with RT-QPCR substantiate the validity of the array data and of the biological conclusions drawn from this study.

Sex-Associated Transcriptomic Profiles
PCA analysis of the gonad gene expression profiles clustered all individuals according to sex and identified marked differences between male and female transcriptomes (Fig. 2). This concurs with previous studies in zebrafish and other vertebrates, where large numbers of transcripts in the gonads are either differentially regulated in males compared with females, or transcripts are sex specific (mouse: Refs. 38, 42; zebrafish: Refs. 20, 49; Drosophila: Ref. 34).


Figure 2
View larger version (32K):
[in this window]
[in a new window]

  		Fig. 2. Analysis of the role of gender as a determinant of gonadal gene expression. Individual males are represented in blue and individual females in red. A: principal component analysis (PCA) of the genes expressed in the gonads (8,769 genes) showing that individual gonad transcriptomes cluster according to sex [x-axis: PCA component 1 (41.44% variance); y-axis: PCA component 2 (15.16% variance)]. B: cluster diagram of the genes differentially expressed between sexes in the gonads (gene tree is displayed horizontally and condition tree is displayed vertically, where males are represented by the blue vertical lines and females by the red vertical lines). The list of genes was generated by comparing male and female gonad transcriptomes (t-test, P < 0.01), with multitesting correction (Benjamini and Hochberg false discovery rate; 1,370 genes). Clustering for both genes and conditions was performed using distance as similarity measure.

 
Statistical analysis of the genes differentially expressed between males and females identified a list of 2,940 genes (t-test, P < 0.05), of which 1,570 were overexpressed in females and 1,370 were overexpressed in males (Fig. 2; additional files 1 and 2). (The online version of this article contains supplemental material.) GO analysis identified 11 biological processes, 3 molecular functions, and 8 cellular components overrepresented in the list of genes overexpressed in females, and 6 biological processes, 5 molecular functions and 9 cellular components overrepresented in the list of genes overexpressed in males (Fig. 3).


Figure 3
View larger version (17K):
[in this window]
[in a new window]

  		Fig. 3. Gene Ontology (GO) analysis of the list of overexpressed genes in ovaries and testes. Data are presented as P value of the overrepresentation of each GO term in the gene lists of interest. A: biological processes represented in the list of genes overexpressed in the ovaries (white bars) and biological processes represented in the list of genes overexpressed in the testes (black bars): 1, biological process unknown; 2, cell communication; 3, cell-cell signaling; 4, cell recognition; 5, host-pathogen interaction; 6, response to endogenous stimulus; 7, response to external stimulus; 8, response to abiotic stimulus; 9, response to biotic stimulus; 10, signal transduction; 11, cell growth and/or maintenance; 12, cell cycle; 13, cell growth; 14, cell homeostasis; 15, cell organization and biogenesis; 16, cytoplasm organization and biogenesis; 17, organelle organization and biogenesis; 18, cytoskeleton organization and biogenesis; 19, mitochondrion organization and biogenesis; 20, cell proliferation; 21, chemi-mechanical coupling; 22, metabolism; 23, biosynthesis; 24, carbohydrate metabolism; 25, coenzymes and prosthetic group metabolism; 26, electron transport; 27, energy pathways; 28, lipid metabolism; 29, nucleobase, nucleoside, nucleotide and nucleic acid metabolism; 30, DNA metabolism; 31, transcription; 32, protein metabolism; 33, protein biosynthesis; 34, protein modification; 35, response to stress; 36, transport; 37, ion transport; 38, protein transport; 39, death; 40, cell death; 41, development; 42, cell differentiation; 43, embryonic development; 44, growth; 45, morphogenesis; 46, regulation of gene expression, epigenetic; 47, reproduction; 48, physiological processes; 49, viral life cycle. B: cellular components represented in the list of genes overexpressed in the ovaries (white bars); and cellular components represented in the list of genes overexpressed in the testes (black bars). 1, cell; 2, intracellular; 3, chromosome; 4, cytoplasm; 5, cytoplasmic chromosome; 6, cytoplasmic vesicle; 7, cytoskeleton; 8, cytosol; 9, endoplasmic reticulum; 10, endosome; 11, Golgi apparatus; 12, microtubule organizing center; 13, mitochondrion; 14, peroxisome; 15, ribosome; 16, vacuole; 17, lysosome; 18, nucleus; 19, nuclear chromosome; 20, nuclear membrane; 21, nucleolus; 22, nucleoplasm; 23, plasma membrane; 24, cellular component unknown; 25, extracellular; 26, extracellular matrix; 27, extracellular space. C: molecular functions represented in the list of genes overexpressed in the ovaries (white bars); and molecular functions represented in the list of genes overexpressed in the testis (black bars). 1, antioxidant activity; 2, apoptosis regulator activity 3, binding; 4, calcium ion binding; 5, carbohydrate binding; 6, lipid binding; 7, nucleic acid binding; 8, DNA binding; 9, chromatin binding; 10, transcription factor activity; 11, nuclease activity; 12, RNA binding; 13, translation factor activity, nucleic acid binding; 14, nucleotide binding; 15, oxygen binding; 16, protein binding; 17, cytoskeletal protein binding; 18, actin binding; 19, catalytic activity; 20, hydrolase; 21, peptidase activity; 22, protein phosphatase activity; 23, kinase; 24, protein kinase activity; 25, transferase; 26, cell adhesion molecule activity; 27, chaperone activity; 28, defense immunity protein activity; 29, molecular function unknown; 30, motor activity; 31, protein tagging activity; 32, signal transducer activity; 33, receptor activity; 34, receptor binding; 35, structural molecule activity; 36, transcription regulator activity; 37, transporter activity; 38, electron transporter activity; 39, ion channel activity; 40, neurotransmitter transporter activity.

 
Among the genes overexpressed in ovaries compared with testes, the biological processes overrepresented included several categories associated with biosynthesis and metabolism. This likely reflects the rapid rate of follicle cell division and the active synthesis and processing of proteins within the oocyte. Overrepresentation of these categories in ovaries has also been observed in Drosophila (34). Among the cellular components associated with this list, endoplasmic reticulum, Golgi apparatus, and cytoplasm were overrepresented, further indicating the importance of protein synthesis and modification in the rapid growth of the oocytes. In addition, the biological processes cell cycle, cell growth and maintenance, and biogenesis of several cellular organelles were also overrepresented in the list of genes overexpressed in the ovary, reflecting the rapid dynamics of cell division, growth, and turnover in the ovary. Among the genes overexpressed in the testes compared with ovaries, the biological processes overrepresented were dominated by protein biosynthesis and protein metabolism. Furthermore, the cellular component most overrepresented was ribosome and the molecular function most overrepresented was transferase, reflecting a significant bias toward protein synthesis in testes; this is consistent with sperm being continuously produced in very large numbers within the testes and released daily in this species.

Expression Profiles of Genes Related to Reproduction
Having established that the gene expression profiles vary considerably between males and females at the level of the gonad, we explored the dataset for sex-related differences in the profiles of genes of known involvement with reproductive function. In these comparisons we included estrogen receptors, insulin-like growth factors (IGFs) and their receptors, genes related to the activin-inhibin pathway, and the sox family of transcription factors. Comparison of the expression profiles of individual genes indicated that there was sex-related differences in the expression profiles of some sox genes, activins, and insulin-like growth factors, and they are presented in Fig. 4.


Figure 4
View larger version (36K):
[in this window]
[in a new window]

  		Fig. 4. Gene expression profiles of families of genes with known involvement in reproductive function. Gene trees are displayed horizontally and were generated using smooth correlation as a similarity measure. Fish 1–5 are males, and fish 7–12 are females. *Statistical significance (P < 0.05).

 
Sox Family
The sox gene family is characterized by sequence homology to the HMG box region of the sex determining gene sry, although there is no functional link between sry and members of the sox family of transcription factors. Several members of the sox gene family have been reported to be important for development of the central nervous system during embryogenesis and of maternal origin, including sox11b and sox21 (37) and sox31 (13). In our study, sox 11b, sox21a, and sox31 were accordingly overexpressed in ovaries compared with testes (P < 0.05). sox19a was also overexpressed in ovaries compared with testes (P < 0.05), but the functional significance of this is not known. sox9 has been proposed as an important regulator of gonadal sex determination in vertebrates (reviewed in Ref. 28). The expression of sox9 in mammals is followed by the expression of anti-Müllerian hormone (amh), leading to inhibition of the peak of aromatase A and regression of the Müllerian ducts, resulting in masculinization of the undifferentiated gonads. In contrast, in the absence of sry, an elevated sox9 gene expression does not occur, and this leads to ovarian development (reviewed in Ref. 19). In the zebrafish, sox9a has been reported to be predominantly expressed in testes (together with amh) and sox9b predominantly expressed in ovaries [together with aromatase A (40)]. In our array dataset, sox9b was found to follow the expected trend, but the difference in gene expression was not statistically significant. sox9a did not pass the quality thresholds for our array data analyses. The functional role of amh in fish is not fully known, but it has been shown to be overexpressed in the testes of zebrafish during early development (Schultz R, unpublished observations) and in adults (40). Similarly, our study revealed that amh and its receptor were greatly overexpressed in the zebrafish testis compared with the ovary (Fig. 5). Together, these findings support the functional involvement of these genes, not only in sex differentiation, but possibly also in the maintenance of sexual dimorphism in adult zebrafish.


Figure 5
View larger version (11K):
[in this window]
[in a new window]

  		Fig. 5. RT-QPCR analysis of 10 transcripts differentially expressed between ovaries and testes. Data are presented as fold difference in relative mRNA expression against the sex with the lower expression. Relative mRNA expression was determined as the ratio of target gene mRNA/rpl8 mRNA. The genes overexpressed in males were anti-Müllerian hormone (amh), tubulin alpha 7 (tuba7), septin 4 (sept4), heat shock protein 70 (hsp70), and cyclin g2 (ccng2). The genes overexpressed in females were RNA binding protein with multiple splicing 2 (rbpms2), transmembrane phosphatase with tensin homology (tpte), sox11b, cyclin b2 (ccnb2), and connexin 44.2 (cx44.2).

 
Activin-Inhibin Pathway
Inhibin exists in two forms, sharing the same {alpha}-subunit and, when covalently linked to one of two distinct subunits called ß-A and ß-B, strongly inhibits pituitary secretion of follicle stimulating hormone (FSH). Dimers of two ß-subunits, however, termed activin, are potent stimulators of FSH secretion (reviewed in Ref. 14). These ß-subunits share extensive sequence homology with transforming growth factor (TGF)-ß and, in addition to their function in regulating reproduction, they also play important roles in development (8), reviewed in Ref. 5. Proposed roles for activin ß-A include promoting ovary and follicle growth, whereas activin ß-B has a tonic role throughout follicle development but becomes critical at the late stage of oocyte maturation and/or ovulation (48). Inhibin ß-B and activin receptor 2B were overexpressed in the testes compared with ovaries (P = 0.008 and P = 0.017, respectively), and this is consistent with this pathway being activated in mature testes that are continuously producing gametes. Forkhead box H1 (FoxH1), a member of the FOX gene family of transcription factors, takes part in mediating TGF-ß/activin signaling through its interaction with the Smad2.Smad4 complex (2). In addition, FoxH1 is a co-repressor of the androgen receptor (6). In all organisms examined, FoxH1 is expressed primarily during the earliest stages of development and thus FoxH1 is thought to play a critical role in mediating TGF-ß superfamily signals during these early developmental stages (2) and is expressed maternally in Xenopus embryos (18). In our study on the zebrafish, foxh1 was overexpressed in ovaries (P = 0.02), reflecting the presence of mature oocytes within the ovaries in all females, where maternal transcripts important for embryo development are synthesized and stored. In addition, the overexpression of foxh1 in the ovaries may be related to the suppression of the biological activity of androgens in this organ.

IGFs
In vertebrates, growth and reproduction are endocrinologically interlinked (reviewed in Refs. 15, 16). The growth axis is centrally regulated by growth hormone, the IGF family, their receptors, and binding proteins. In addition, these genes also regulate development (reviewed in Refs. 41, 44) and male sex determination (32). At the level of the gonad, IGFs play central roles in mediating gonadotropin function (54), regulating steroidogenesis (25, 29) and promoting oocyte growth and maturation (29, 35). In the zebrafish, genes for two IGFs have been isolated (IGF1 and IGF2, reviewed in Ref. 51); igf1 is predominantly expressed in the liver and testes, whereas igf2 is expressed predominantly in the liver (27). Similarly, in our studies igf1 appeared to be overexpressed in testis (igf2 did not pass the quality thresholds in our analysis, possibly due to its low expression in the gonads). Two forms of the IGF1 receptor are expressed in the zebrafish [insulin-like growth factor 1a receptor (igf1ra), insulin-like growth factor 1b receptor (igf1rb)], and they were expressed similarly in the ovaries and testes in the study by Maures and colleagues (27). In our study, igf1ra was expressed equally in the gonads of both sexes, but igf1rb was predominantly expressed in the ovary (P = 0.033). Very recently, studies in the fathead minnow using RT-QPCR (9) showed comparable results to our array dataset for igf1r in the zebrafish, but different roles for igf1ra and igf1rb have yet to be established. The endocrine and paracrine functions of IGFs are further regulated by IGF binding proteins (IGFBP). IGFBPs protect IGFs from degradation and regulate their bioavailability, preventing hypoglycemia caused by excess of IGFs in the circulation. In addition, they play a role during development of many organs in the zebrafish, as demonstrated by Li and colleagues (21) and Wood and colleagues (52). To date, six IGFBPs have been reported in fish, and their expression was localized to a number of tissues including the liver and ovaries (17). In our studies, igfbp3 was the only form consistently expressed in both male and female gonads, and its expression was similar in both sexes. Our findings indicate that genes for all constituents required for the biological activity of IGFs (igf1, igfr, and igfbp) were expressed in the gonads of both males and females, reflecting their paracrine mode of action. The differential expression pattern for igf1rb suggests distinct regulations of IGF function in male and female zebrafish gonads.

Sex-specific Genes
Among the genes differentially expressed between ovaries and testes in the array dataset, 52 genes showed >10-fold difference in expression between sexes. A selection of differentially expressed genes from the array analyses were further analyzed using RT-QPCR (Fig. 5), and the patterns of differential expression of these transcripts were confirmed. There was up to approximately a 3,000-fold difference in expression of some of these genes between ovaries and testes. The magnitude of the differences in expression between the sexes suggests a functional silencing of these genes in one of the sexes. These genes are therefore likely to play crucial roles in the context of sex-specific gonad function. Further studies are required to explore the functional significance of these findings at the level of the individual genes.

Anchoring Individual Transcriptomes With Reproductive Phenotype
The gene expression profiles of the testis were remarkably similar between mature males with the correlation coefficients ranging from 0.481 to 0.685 when the transcriptome of each individual was compared with all other males. Analysis of variance (ANOVA) failed to identify any genes differing significantly between individual males (Fig. 6). This is not necessarily surprising given that all stages of spermatogenesis including spermatozoa were present in the testes of all males. However, very significant differences in sperm quality between individual males within a breeding colony were observed, and both the number and motility of spermatozoa varied considerably (46). Based on these observations, we hypothesized that variation in the expression of small sets of genes within the testes might account for the phenotypic differences in sperm quality observed between individual males. We searched for such possible differences using three independent statistical analyses of the microarray data. Firstly, we investigated genes varying more than twofold between any two individuals, and this generated a list of 395 genes. Secondly, we assigned individual males into categories according to the similarity of their overall transcriptome (based on cluster analysis of all consistently expressed genes) and performed comparisons by the t-test (when the fish were assigned to one of two groups) or one-way ANOVA (when the fish were assigned to one of three groups), using the Benjamini and Hochberg false discovery rate as multiple testing corrections. This approach did not identify any differentially expressed genes. Without multiple testing corrections, lists of 228 and 209 genes were identified using the t-test and one-way ANOVA, respectively. When the three lists generated were compared, no common genes were identified and we concluded that the approaches used likely identified the highly variable genes rather than genes encoding for physiologically significant testicular proteins. Given the consistency in the testis transcriptome between individuals, the differences in sperm motility may be due to posttranscriptional events (such as translation and protein modifications) rather than large differences in gene expression. In humans, altered patterns of expression have been identified for testis-specific protein 1 and lactate dehydrogenase C transcript variant 1 in males with reduced fertility (47). In our study, however, we do not know whether the observed differences in sperm quality translated into differences in fertilization capability. Our data generate the hypothesis that the transcripts required for sperm function are synthesized and stored in spermatozoa, possibly during earlier stages of germ cell development, and that the progression of individual spermatozoa towards motile sperm depends on the subsequent activation of the resulting gene products.


Figure 6
View larger version (19K):
[in this window]
[in a new window]

  		Fig. 6. Comparisons of the gene expression profiles between individual males and individual females in a zebrafish breeding colony. A: correlation coefficients of the transcriptome of individual males compared with all other males. B: correlation coefficients of the transcriptome of individual females compared with all other females. C: Venn diagram of the genes differentially expressed between individual males [a, genes up- or downregulated by >2-fold in at least 2 individuals; b, genes differentially expressed between clusters when individual males were assigned to 3 groups (ANOVA); c, genes differentially expressed between clusters when individual males were assigned to 2 groups (t-test)]; D: Venn diagram of the genes differentially expressed between individual females [a, genes up- or downregulated by >2-fold in at least 2 individuals; b, genes differentially expressed between clusters when individual females were assigned to 3 groups, according to the histopathology of the gonads (ANOVA); c, genes differentially expressed between clusters when individual females were assigned to 2 groups according to the cluster analysis of all consistently expressed genes (t-test)].

 
The variation between the gonadal transcriptomes of individual females exceeded that observed in males. The correlation coefficients, when the gonadal transcriptomic profiles of each female are compared with the transcriptomic profile of all other females, ranged between 0.198 and 0.461 (Fig. 6). Individual females were grouped into two categories according to the condition tree generated using all expressed genes (using distance as similarity measure). The first category was characterized by predominance of earlier stages of oogenesis, and in the second category vitellogenic follicles were most abundant. Females were also classified into three categories based on ovarian histology, in which the first category comprised one female where primary oocytes were predominant, the second comprised females where vitellogenic oocytes were predominant, and the third comprised one female displaying large batches of postovulatory oocytes, indicating that this individual had recently undergone a significant spawning event. Comparisons of the gene expression profiles for females in each of these groups generated lists of differentially expressed genes, which had a relatively high level of overlap with the genes changing more than twofold between individual females (Fig. 6). GO analysis of the genes differentially expressed between each of the groups identified a number of overrepresented biological processes: in females where earlier stages of oogenesis were predominant, among the genes overexpressed, the GO terms overrepresented were protein metabolism, protein biosynthesis, metabolism, and cell growth and/or maintenance, and this is consistent with rapid rates of growth and development of the ovarian follicles at this stage. Within maturing ovaries, follicle growth comprises two main events: in the first place the oocyte within the follicle undergoes large increases in size mainly due to the incorporation of exogenous proteins, such as vitellogenin. This requires batteries of enzymes (such as cathepsins) able to process this large yolk precursor into smaller molecules, which are stored within the oocyte to provide nutritional support for the developing embryos (reviewed in Ref. 45). Secondly, follicle growth requires the division and specialization of somatic cells to produce the layers surrounding the oocytes and providing both hormonal and nutritional support for the developing oocytes (reviewed in Ref. 30). The biological processes overrepresented in maturing females are, therefore, indicative of this fast phase of growth of the ovarian follicles undergoing secondary growth. In contrast, in females where mature oocytes were predominant, the biological processes overrepresented among the genes overexpressed were embryonic development, development, and cell growth and/or maintenance. This reflects the transcription and storage of genes of maternal origin that support embryonic development before the activation of the zygotic genome [and to a lesser extent also after this event (26)]. In the female where large batches of postovulatory follicles were evident, a single biological process was overrepresented (among the genes underexpressed in this individual compared with the other females in the breeding colony) and this was response to stress. The physiological significance of this finding is not clear; however, exposure to stress and the ability to cope with stress are likely to be important factors in the number and quality of the gametes produced. Studies in the rainbow trout revealed that exposure to stress during oogenesis resulted in lower quality of the gametes produced (4, 7). Further studies are required to fully establish the implications of exposure to stress on reproductive success.

Cluster analysis based on the 13 genes found to overlap between the three methods used to compare individual ovaries (Fig. 6D) grouped females into two groups (the first group was characterized by predominance of earlier stages of oogenesis, and the second group by a predominance of vitellogenic follicles). The expression patterns of the genes identified were consistent with the molecular events characteristic of oocyte growth and maturation. Among the genes overexpressed in the first group, we identified synaptonemal complex protein 1, a gene involved in maintaining the interaction in recombinant chromosomes during prophase I of the meiotic division in germ cells (33). This reflects the fact that most germ cells contained in these ovaries were arrested at prophase I of the meiotic division (the early stages of oocyte growth in teleosts occur during this phase). In ovaries containing large numbers of vitellogenic follicles, integral membrane protein 2b, a gene involved in inducing apoptosis (12), and transducer of ERBB2, 1a, a gene that displays antiproliferative properties (53) were overexpressed. These profiles relate to the high proportion of follicles in mature females that are at a stage close to ovulation and characterized by an arrest in follicle cell division.

In conclusion, the present data set provides the first detailed description and analysis of the reproductive gonadal transcriptome of actively breeding zebrafish at the individual level.

Large sex-related differences in gene expression were identified between ovaries and testes of breeding fish. Many novel transcripts showing marked differential expression between genders were identified allowing for new insights into the molecular mechanisms controlling reproduction in this model species. Relationships between the ovarian transcriptome and reproductive status in females were also established.


   GRANTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 GRANTS
REFERENCES
 
This work was funded by Biotechnology and Biological Sciences Research Council Grant 9/S15001 and Natural Environment Research Council (UK) Environmental Genomics Thematic Grant NER/T/S/2002/00182 to C. R. Tyler.


   ACKNOWLEDGMENTS
 
The authors thank members of the Environmental and Molecular Fish Biology Group at the University of Exeter for technical support and the Wales Gene Park and its staff at Cardiff for access to their facilities and technical support. We also thank Dr. R. van Aerle for critical comments and technical editing of the images for this manuscript.


   FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: E. M. Santos, School of Biosciences, Univ. of Exeter, Prince of Wales Rd., Exeter, EX4 4PS, UK (e-mail: E.Santos{at}exeter.ac.uk).

* P. Kille and C. R. Tyler are co-senior authors and contributed equally to this work. Back


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 GRANTS
 REFERENCES
 

  1. Ahluwalia A, Clodfelter KH, Waxman DJ. Sexual dimorphism of rat liver gene expression: Regulatory role of growth hormone revealed by deoxyribonucleic acid microarray analysis. Mol Endocrinol 18: 747–760, 2004.[Abstract/Free Full Text]
  2. Attisano L, Silvestri C, Izzi L, Labbe E. The transcriptional role of Smads and FAST (FoxH1) in TGFbeta and activin signalling. Mol Cell Endocrinol 180: 3–11, 2001.[CrossRef][ISI][Medline]
  3. Blazquez M, Bosma PT, Fraser EJ, Van Look KJW, Trudeau VL. Fish as models for the neuroendocrine regulation of reproduction and growth. Comp Biochem Physiol C 119: 345–364, 1998.[ISI][Medline]
  4. Campbell PM, Pottinger TG, Sumpter JP. Stress reduces the quality of gametes produced by rainbow trout. Biol Reprod 47: 1140–1150, 1992.[Abstract]
  5. Chang H, Lau AL, Matzuk MM. Studying TGF-beta superfamily signaling by knockouts and knockins. Mol Cell Endocrinol 180: 39–46, 2001.[CrossRef][ISI][Medline]
  6. Chen G, Nomura M, Morinaga H, Matsubara E, Okabe T, Goto K, Yanase T, Zheng H, Lu J, Nawata H. Modulation of androgen receptor transactivation by FoxH1. A newly identified androgen receptor corepressor. J Biol Chem 280: 36355–36363, 2005.[Abstract/Free Full Text]
  7. Contreras-Sanchez WM, Schreck CB, Fitzpatrick MS, Pereira CB. Effects of stress on the reproductive performance of rainbow trout (Oncorhynchus mykiss). Biol Reprod 58: 439–447, 1998.[Abstract/Free Full Text]
  8. Dudas M, Kaartinen V. Tgf-beta superfamily and mouse craniofacial development: interplay of morphogenetic proteins and receptor signaling controls normal formation of the face. Curr Top Dev Biol 66: 65–133, 2005.[CrossRef][ISI][Medline]
  9. Filby AL, Thorpe KL, Tyler CR. Multiple molecular effect pathways of an environmental oestrogen in fish. J Mol Endocrinol 37: 121–134, 2006.[Abstract/Free Full Text]
  10. Filby AL, Tyler CR. Appropriate 'housekeeping' genes for use in expression profiling the effects of environmental estrogens in fish. BMC Mol Biol 8: 10, 2007.[CrossRef][Medline]
  11. Filby AL, Tyler CR. Molecular characterization of estrogen receptors 1, 2a, and 2b and their tissue and ontogenic expression profiles in fathead minnow (Pimephales promelas). Biol Reprod 73: 648–662, 2005.[Abstract/Free Full Text]
  12. Fleischer A, Ayllon V, Rebollo A. ITM2B(S) regulates apoptosis by inducing loss of mitochondrial membrane potential. Eur J Immunol 32: 3498–3505, 2002.[CrossRef][ISI][Medline]
  13. Girard F, Cremazy F, Berta P, Renucci A. Expression pattern of the Sox31 gene during zebrafish embryonic development. Mech Dev 100: 71–73, 2001.[CrossRef][ISI][Medline]
  14. Gregory SJ, Kaiser UB. Regulation of gonadotropins by inhibin and activin. Semin Reprod Med 22: 253–267, 2004.[CrossRef][ISI][Medline]
  15. Hull KL, Harvey S. Growth hormone: roles in female reproduction. J Endocrinol 168: 1–23, 2001.[Abstract]
  16. Hull KL, Harvey S. Growth hormone: roles in male reproduction. Endocrine 13: 243–250, 2000.[CrossRef][ISI][Medline]
  17. Kamangar BB, Gabillard JC, Bobe J. Insulin-like growth factor-binding protein (IGFBP)-1,-2,-3,-4,-5, and -6 and IGFBP-related protein 1 during rainbow trout postvitellogenesis and oocyte maturation: molecular characterization, expression profiles, and hormonal regulation. Endocrinology 147: 2399–2410, 2006.[Abstract/Free Full Text]
  18. Kofron M, Puck H, Standley H, Wylie C, Old R, Whitman M, Heasman J. New roles for FoxH1 in patterning the early embryo. Development 131: 5065–5078, 2004.[Abstract/Free Full Text]
  19. Koopman P. Sry and Sox9: mammalian testis-determining genes. Cell Mol Life Sci 55: 839–856, 1999.[ISI][Medline]
  20. Li Y, Chia JM, Bartfai R, Christoffels A, Yue GH, Ding K, Ho MY, Hill JA, Stupka E, Orban L. Comparative analysis of the testis and ovary transcriptomes in zebrafish by combining experimental and computational tools. Comp Funct Genomics 5: 403–418, 2004.[CrossRef]
  21. Li Y, Xiang J, Duan C. Insulin-like growth factor-binding protein-3 plays an important role in regulating pharyngeal skeleton and inner ear formation and differentiation. J Biol Chem 280: 3613–3620, 2005.[Abstract/Free Full Text]
  22. Linney E, Dobbs-McAuliffe B, Sajadi H, Malek RL. Microarray gene expression profiling during the segmentation phase of zebrafish development. Comp Biochem Physiol C 138: 351–362, 2004.[ISI]
  23. Lo J, Lee SC, Xu M, Liu F, Ruan H, Eun A, He YW, Ma WP, Wang WF, Wen ZL, Peng JR. 15,000 unique zebrafish EST clusters and their future use in microarray for profiling gene expression patterns during embryogenesis. Genome Res 13: 455–466, 2003.[Abstract/Free Full Text]
  24. Malek RL, Sajadi H, Abraham J, Grundy MA, Gerhard GS. The effects of temperature reduction on gene expression and oxidative stress in skeletal muscle from adult zebrafish. Comp Biochem Physiol C 138: 363–373, 2004.[ISI]
  25. Manna PR, Chandrala SP, King SR, Jo Y, Counis R, Huhtaniemi IT, Stocco DM. Molecular mechanisms of insulin-like growth factor-I mediated regulation of the steroidogenic acute regulatory protein in mouse Leydig cells. Mol Endocrinol 20: 362–378, 2006.[Abstract/Free Full Text]
  26. Mathavan S, Lee SG, Mak A, Miller LD, Murthy KR, Govindarajan KR, Tong Y, Wu YL, Lam SH, Yang H, Ruan Y, Korzh V, Gong Z, Liu ET, Lufkin T. Transcriptome analysis of zebrafish embryogenesis using microarrays. PLoS Genet 1: 260–276, 2005.[Medline]
  27. Maures T, Chan SJ, Xu B, Sun H, Ding J, Duan C. Structural, biochemical, and expression analysis of two distinct insulin-like growth factor I receptors and their ligands in zebrafish. Endocrinology 143: 1858–1871, 2002.[Abstract/Free Full Text]
  28. Morrish BC, Sinclair AH. Vertebrate sex determination: many means to an end. Reproduction 124: 447–457, 2002.[Abstract]
  29. Mukhejee D, Mukherjee D, Sen U, Paul S, Bhattacharyya SP. In vitro effects of insulin-like growth factors and insulin on oocyte maturation and maturation-inducing steroid production in ovarian follicles of common carp, Cyprinus carpio. Comp Biochem Physiol A 144: 63–77, 2006.[CrossRef][Medline]
  30. Nagahama Y, Yoshikuni M, Yamashita M, Tokumoto T, Katsu Y. Regulation of oocyte growth and maturation in fish. Curr Top Dev Biol 30: 103–145, 1995.[ISI][Medline]
  31. Nash JP, Kime DE, Van der Ven LTM, Wester PW, Brion F, Maack G, Stahlschmidt-Allner P, Tyler CR. Long-term exposure to environmental concentrations of the pharmaceutical ethynylestradiol causes reproductive failure in fish. Environ Health Perspect 112: 1725–1733, 2004.[ISI][Medline]
  32. Nef S, Verma-Kurvari S, Merenmies J, Vassalli JD, Efstratiadis A, Accili D, Parada LF. Testis determination requires insulin receptor family function in mice. Nature 426: 291–295, 2003.[CrossRef][Medline]
  33. Paredes A, Garcia-Rudaz C, Kerr B, Tapia V, Dissen GA, Costa ME, Cornea A, Ojeda SR. Loss of synaptonemal complex protein-1, a synaptonemal complex protein, contributes to the initiation of follicular assembly in the developing rat ovary. Endocrinology 146: 5267–5277, 2005.[Abstract/Free Full Text]
  34. Parisi M, Nuttall R, Edwards P, Minor J, Naiman D, Lu JN, Doctolero M, Vainer M, Chan C, Malley J, Eastman S, Oliver B. A survey of ovary-, testis-, and soma-biased gene expression in Drosophila melanogaster adults. Genome Biol 5: 2004.
  35. Patino R, Yoshizaki G, Thomas P, Kagawa H. Gonadotropic control of ovarian follicle maturation: the two-stage concept and its mechanisms. Comp Biochem Physiol B Biochem Mol Biol 129: 427–439, 2001.[CrossRef][Medline]
  36. Rasmussen R. Quantification on the LightCycler. Heidelberg: Springer Press, 2001, p. 21–34.
  37. Rimini R, Beltrame M, Argenton F, Szymczak D, Cotelli F, Bianchi ME. Expression patterns of zebrafish sox11A, sox11B and sox21. Mech Dev 89: 167–171, 1999.[CrossRef][ISI][Medline]
  38. Rinn JL, Rozowsky JS, Laurenzi IJ, Petersen PH, Zou KY, Zhong WM, Gerstein M, Snyder M. Major molecular differences between mammalian sexes are involved in drug metabolism and renal function. Dev Cell 6: 791–800, 2004.[CrossRef][ISI][Medline]
  39. Rinn JL, Snyder M. Sexual dimorphism in mammalian gene expression. Trends Genet 21: 298–305, 2005.[CrossRef][ISI][Medline]
  40. Rodriguez-Mari A, Yan YL, Bremiller RA, Wilson C, Canestro C, Postlethwait JH. Characterization and expression pattern of zebrafish Anti-Mullerian hormone (Amh) relative to sox9a, sox9b, and cyp19a1a, during gonad development. Gene Expr Patterns 5: 655–667, 2005.[CrossRef][Medline]
  41. Silha JV, Murphy LJ. Insulin-like growth factor binding proteins in development. Adv Exp Med Biol 567: 55–89, 2005.[ISI][Medline]
  42. Small CL, Shima JE, Uzumcu M, Skinner MK, Griswold MD. Profiling gene expression during the differentiation and development of the murine embryonic gonad. Biol Reprod 72: 492–501, 2005.[Abstract/Free Full Text]
  43. Spence R, Smith C. Mating preference of female zebrafish, Danio rerio, in relation to male dominance. Behav Ecol 17: 779–783, 2006.[Abstract/Free Full Text]
  44. Trejo JL, Carro E, Burks DJ. Experimental models for understanding the role of insulin-like growth factor-I and its receptor during development. Adv Exp Med Biol 567: 27–53, 2005.[ISI][Medline]
  45. Tyler CR, Sumpter JP. Oocyte growth and development in teleosts. Rev Fish Biol Fish 6: 287–318, 1996.[CrossRef]
  46. Van Look KJW, Paull GC, Santos EM, Tyler CR, Holt WV. Testicular asymmetry in zebrafish. Hum Fertil 9: 107, 2006.[CrossRef]
  47. Wang H, Zhou Z, Xu M, Li J, Xiao J, Xu ZY, Sha J. A spermatogenesis-related gene expression profile in human spermatozoa and its potential clinical applications. J Mol Med 82: 317–324, 2004.[CrossRef][ISI][Medline]
  48. Wang Y, Ge W. Developmental profiles of activin betaA, betaB, and follistatin expression in the zebrafish ovary: evidence for their differential roles during sexual maturation and ovulatory cycle. Biol Reprod 71: 2056–2064, 2004.[Abstract/Free Full Text]
  49. Wen CM, Zhang ZH, Ma WP, Xu M, Wen ZL, Peng JR. Genome-wide identification of female enriched genes in zebrafish. Dev Dyn 232: 171–179, 2005.[CrossRef][ISI][Medline]
  50. Wettenhall JM, Smyth GK. limmaGUI: a graphical user interface for linear modeling of microarray data. Bioinformatics 20: 3705–3706, 2004.[Abstract/Free Full Text]
  51. Wood AW, Duan C, Bern HA. Insulin-like growth factor signaling in fish. Int Rev Cytol 243: 215–285, 2005.[CrossRef][ISI][Medline]
  52. Wood AW, Schlueter PJ, Duan C. Targeted knockdown of insulin-like growth factor binding protein-2 disrupts cardiovascular development in zebrafish embryos. Mol Endocrinol 19: 1024–1034, 2005.[Abstract/Free Full Text]
  53. Xiong B, Rui YN, Zhang M, Shi KH, Jia SJ, Tian T, Yin K, Huang HZ, Lin SY, Zhao XG, Chen YH, Chen YG, Lin SC, Meng AM. Tob1 controls dorsal development of zebrafish embryos by antagonizing maternal beta-catenin transcriptional activity. Dev Cell 11: 225–238, 2006.[CrossRef][ISI][Medline]
  54. Yu Y, Li W, Han Z, Luo M, Chang Z, Tan J. The effect of follicle-stimulating horm