Background
The mechanism underlying sex determination in zebrafish is largely unknown. Currently, no morphological differences in the chromosomes of the two sexes have been identified neither by classical karyotyping nor from the zebrafish genome sequencing project. This indicates that sex determination in zebrafish is mediated by genetic signals from autosomal genes and not by XY/XX or ZZ/ZW sex determination [
1‐
3]. Furthermore, the mammalian sex determining factor
Sry (located on the Y chromosome) and the sex determining gene
dmy (
dmrt1y) located on the Y chromosome in Japanese medaka
Oryzias latipes [
4,
5] does not appear to be present in the zebrafish genome. Current knowledge on sex determination and differentiation in zebrafish includes that zebrafish are sexually mature after three months, and that separate sexes can be detected by gonad histology at approximately 40 days post hatch (dph). Furthermore, all zebrafish develop ovary-like gonads regardless of genetic background prior to sex differentiation [
3]. Ovarian development is initiated at approximately 10 dph and progresses until 20 dph. At 20 dph until approximately 30 dph testis development is initiated in males simultaneously with ovarian cell apoptosis [
2,
3].
A number of genes (
dmrt1,
sox9a,
amh,
wt1,
ftz-f1, gata) have previously been associated with the process of sex determination or differentiation in zebrafish [
2]. None of these genes are suggested to be the single factor responsible for specifying sex in zebrafish [
2]. However, the expression pattern and function of these genes suggest that they are part of a signalling network responsible for the development of sex specific gonads [
2]. Current research in medaka
Oryzias latipes suggests a potential signalling pathway in XY individuals of:
dmy,
sox9a and
dmrt1 with
dmy blocking meiosis,
sox9a regulating testicular tubule development and
dmrt1 being important in spermatogenesis [
6]. Also, the research in medaka
Oryzias latipes suggests an important role of
fig α as well as
cyp19a1a and
cyp19a1b in sexual development of gonads in XX individuals [
7].
The
Dmrt1 gene belongs to a group of multiple genes containing a zinc-finger-like DNA-binding motif (DM domain) that has been identified in both invertebrates and vertebrates [
8]. Interestingly, the medaka sex determining gene
DMY (
dmrt1y) originates from a
dmrt1 duplication, and sex determining homologues have been identified in the fruit fly
Drosophila melanogaster (
doublesex gene) and the worm
Caenorhabditis elegans (
mab-3 gene) [
8]. In zebrafish,
in situ hybridization to gonads showed that
dmrt1 was expressed in developing germ cells of both testis and ovary, suggesting that the
dmrt1 gene is not only associated with testis development, but may also be important in ovary differentiation of zebrafish [
8]. In diverse species including frog, turtle, alligator, bird, and mouse, the
dmrt1 gene is expressed at higher levels in males compared to females, suggesting that high expression is necessary for testicular differentiation, whereas lower expression is compatible with ovarian differentiation [
9‐
12]. The
Sox (
SRY-related genes containing a HMG box) gene family encodes an important group of developmental regulators involved in sex determination. The HMG (high mobility group) box that characterises Sox proteins is a DNA-binding domain and proteins encoded by
Sox genes act as transcription factors.
Sry, the founder member of the
Sox gene family, is the Y chromosomal male determinant in most mammals [
13‐
15]
Sry is a poorly conserved gene that appears to be exclusive to mammals. In contrast,
Sox 9 is a conserved gene present in all vertebrate types. Like
Sry,
Sox 9 is required for testis development in mammals, and
Sox9 deficiency can result in sex reversal in human males. The expression of
Sox9 during gonadal differentiation is up-regulated in testis and down-regulated in ovaries in mammals, birds and turtles [
16]. However, the organisation and function of the
Sox gene family is less understood in other types of vertebrates and despite the wide distribution of
Sox9 genes in fish, only few have been investigated [
16,
17]. In zebrafish, two
sox9 genes are present
sox9a and
sox9b) and their expression patterns indicate that they have unique functions during development [
16]. In adult zebrafish the
sox9a transcript was observed in testis but not in ovary. Conversely,
sox9b transcripts were detected in ovary, but not in testis [
16].
Androgens play a key role in male sex differentiation and sex maturation in vertebrates, including teleosts [
18]. The most important androgens in teleost fish are 11-ketotestosterone and testosterone and their action is mediated through specific nuclear androgen receptors (
ar) [
19‐
21]. In sea bass (
Dicentrarchus labrax) expression of the
ar was measured in the gonads during development of a male-dominated population and a female-dominated population [
22]. The expression patterns of the two populations were different, with a peak in the
ar expression in the male-dominated population coinciding with the time of sex differentiation in sea bass [
22]. We recently identified and conducted the initial characterisation of a novel androgen receptor from zebrafish [
23] and in another recent study, expression of this receptor was higher in the transforming testis compared to ovary suggesting a role during male gonadal differentiation [
24]. A key enzyme balancing the ratio of steroid hormones is cyp19 (aromatase) which is the terminal enzyme in the steroidogenic pathway. It converts androgens (e.g., testosterone) into estrogens (e.g., estradiol). Regulation of this gene dictates the ratio of androgens to estrogens; therefore, appropriate expression of this enzyme is critical for sex differentiation and reproduction in vertebrates. Most vertebrates have a single
CYP19 gene that is regulated by multiple tissue-specific promoter regions. However, the zebrafish has two genes (
cyp19a1a and
cyp19a1b) encoding different proteins and each possessing its own regulatory mechanism [
1]. In general, the gonadal form
cyp19a1a is more abundant than the brain form (
cyp19a1b). The expression of the two
cyp19 genes has previously been investigated from 0–41 days post fertilisation (dpf) which is the expected time of sex determination and differentiation in zebrafish.
cyp19a1a expression was highest shortly after hatch from 4–8 dpf. The pattern of
cyp19a1b expression was segregated into two populations, suggesting an association with sex differentiation [
1].
fig α is a germ cell-specific transcription factor required for ovarian follicle formation.
fig α is involved in the coordinate expression of the zona pellucida (zpc) genes. The expression of
fig α and
zpc coincides with the onset of gonadal differentiation in zebrafish at 22 dpf [
25]. Previous in situ hybridisation studies of
fig α expression in zebrafish have shown that
fig α is expressed abundantly in ovaries of adult fish whereas no
fig α signal could be detected in adult zebrafish testes [
25]. Furthermore, at 30 dpf the expression of
fig α and
zpc was investigated in eight fish. In five fish both genes were expressed whereas no expression was detected in the last three fish. This indicates female restricted expression of
fig α in zebrafish [
25].
Since previous studies indicate involvement of ar, sox9a, dmrt1, cyp19a1b, cyp19a1a and fig α in sex differentiation in vertebrates, including fish, we found it interesting to investigate the expression of these six genes during sex determination and differentiation in zebrafish. The idea is to measure the expression of all genes in the same individual, thereby allowing comparison of expression. This might lead to identification of putative male and female zebrafish and indicate whether one or more of these genes could function as an early genetic sex marker in zebrafish. Therefore, the aim of the current study is to determine the expression of the ar, sox9a, dmrt1, cyp19a1b, cyp19a1a and fig α during the developmental period in which sex determination and gonadal differentiation takes place in zebrafish.
Methods
Animals
Juvenile zebrafish originated from a brood population of fish. In the evening breeding boxes were placed in an aquarium with parent fish and eggs were collected the following morning. The eggs were sorted into fertilised and non-fertilised and the fertilised eggs were placed in 900 ml glass beakers and kept at 26 +/- 1°C and a light-dark period of 14:10 h. In the interval 3–22 dph the larvae were fed two times daily with powdered dry food (Sera Micron) and one time daily with newly hatched artemia sp. nauplii (Intér Ryba GmbH, Germany). For a period of three days (23–25 dph) the powdered dry food was given in combination with TetraMin Baby Powder Food. From 26 dph and onwards the dry food consisted solely of TetraMin Baby Powder Food. Artemia was still given once daily. At 2, 4, 6, 8, 10, 12, 14, 16, 18, 19, 20, 21, 22, 25, 30 and 40 dph zebrafish were frozen individually in liquid nitrogen and stored at -80°C until further analysis.
RNA purification
Total RNA was purified from whole fish for the AR expression during sex determination and differentiation experiment. Zebrafish or zebrafish tissue was homogenised for 20 sec. using an Ultra-Turrax homogenizer (IKA-Werke). For purification of RNA, Total RNA Isolation Kit (Macherey-Nagel) was used according to manufacturer's instructions. RNA concentration was measured spectrophotometrically (Gene-Quant, Pharmacia Biotech), checked by gel electrophoresis (1.2% agarose gel) and stored at -80°C until further use. cDNA was obtained from 0.3 μg RNA using SuperScript II reverse transcriptase kit with Oligo dT primer (Invitrogen) according to manufacturers instructions.
Quantitative RT-PCR
Expression of
Danio rerio ar,
sox9a,
dmrt1,
cyp19a1b,
fig α and
cyp19a1a was analysed by qRT-PCR using a real-time light-cycler (Roche). As zebrafish is an established model-organism, the genome has been sequenced and the seventh assembly is available. All genes investigated have previously been identified and sequences were available in GenBank. Primers for qRT-PCR analysis were designed using the Primer3 program [
26] (Table
1). The final PCR reactions contained: 0.4 mM of each primer; 0.25 × SYBR Green (Invitrogen); 4 mM MgCl
2 and as template 5 μl of cDNA reverse transcribed from a standardized amount of total RNA (0.3 μg). qRT-PCR was performed using Hotstart Taq polymerase (Qiagen) in a final volume of 20 μl. All quantitative reactions were subjected to: 95°C for 15 min followed by 45 cycles at 94°C for 15 s, 59°C 15 s and 72°C 15 s. Melting curve analysis was applied to all reactions to ensure homogeneity of the reaction product. In addition, the amplicon size was checked electrophoretically for each primer set and subsequent sequencing revealed that it corresponded to the zebrafish
ar,
sox9a,
dmrt1,
cyp19a1b, fig α and
cyp19a1a, thus verifying the identity of the genes. Potential contamination was assessed by including non-reverse transcribed total RNA (genomic DNA contamination) and no-template controls. No products were observed in these reactions. Dilution curves generated by serial dilutions (1:10) of cDNA were used to calculate amplification efficiencies according to the Roche protocol. All assays were quantitative with standard curve (mean threshold cycle [C
t] vs. log cDNA dilution) slopes of -3.96 (
β-actin), -2.10 (
sox9a), -3.39 (
dmrt1), -2.88 (
ar), -3.00 (
cyp19a1a), -4.22 (
cyp19a1b) and -3.68 (
fig α), translating to relatively high PCR efficiencies (E) of 1.79 (
β-actin), 3.01 (
sox9a), 1.97 (
dmrt1), 2.23 (
ar), 2.15 (
cyp19a1a), 1.73 (
cyp19a1b) and 1.87 (
fig α). over the detection range, the linear correlation (R
2) between the mean C
t and the log cDNA dilution was > 0.99 in each case. Transcript levels of the target genes were normalized to
D. rerio β-actin after correcting for differences in amplification efficiencies.
Table 1
Oligonucleotide primers used for quantitative real time PCR analysis.
ar | 242 nt | AGCAGCAGCACCACTACCA | TTCCTTCCTGCCTCTCGTTC |
sox9a | 719 nt | CGGTGAAGAACGGCCAGAGC | CTGTAGAGTCAGCAATGGGT |
cyp19a1b | 230 nt | AACATTGGACGCATGCATAA | TGTTTGATGGTGCTGATGGT |
dmrt1a | 151 nt | ATGGCAGAGCAGAACGATTT | TAGTCCCACAACAGCATGGA |
fig α | 663 nt | ATGTCGTGTGAAATGACCGGC | CTAGGATGGGAGTGAACTTGG |
cyp19a1a | 131 nt | AGATGTCGAGTTAAAGATCCTGCA | CGACCGGGTGAAAACGTAGA |
β-actin | 272 nt | CCTGACCGAGAGAGGCTACA | CGCAAGATTCCATACCCAAG |
Tissue expression
Adult zebrafish, 5 males and 5 females were anaesthetized in a buffered solution of MS-222 (0.1 g/l) and quickly dissected into brain, gonads, liver, eyes, spleen, heart, gut, gall bladder, muscle and gills. The dissected tissues were immediately frozen in liquid nitrogen and stored at -80°C. The RNA purification and quantitative qRT-PCR was conducted as described for the gene expression during sex determination and differentiation experiment except that approximately 5 mg of zebrafish tissue were used for total RNA purification. Also, 1 μg total RNA was reverse transcribed using SuperScript II reverse transcriptase kit with Oligo dT primer (Invitrogen) according to manufacturers instructions. Following the qRT-PCR, the reactions were spinned down and loaded on a 0.8% agarose gel.
Data handling and statistical analysis
To analyse the expression patterns of ar, sox9a, dmrt1, fig α, cyp19a1a and cyp19a1b during the investigated period, it was necessary to discriminate between individual fish with high and low gene expression. This was done using the GraphPad Prism program makes a cut-off value between the high and low expressers as indicated on the graph, which is generated automatically in the program when making a two-segment graph. The investigated fish were a mixture of males and females and since large differences in expression were observed during the investigated period, it was difficult to separate the different effects from each other without this distinction between high and low expressers.
As expression of all the six investigated genes has been measured on individual fish, it is possible to compare the expression patterns. First, ar, sox9a and dmrt1 genes with expected high expression in males were divided into high and low expressers, then for each individual fish the expression levels (high or low) of the three genes were compared. Next, the percentage of individuals with high or low expression of the three genes was calculated. Secondly, the same was calculated for the two genes (fig α and cyp19a1a) with expected high expression in females. Third, the percentage of individuals with expression patterns as expected for male or female for all five genes was calculated.
Gene expressions in zebrafish during the sex determination and differentiation period were statistically analysed for variance. Data were normally distributed with similar variance after log transformation and a two-way ANOVA (variables: high/low expressers and dph) analysis was followed by Tukey test with a significance level of (p < 0.05).
Discussion
Testicular type
sox9 is the most upstream conserved gene in the sex determining cascade among vertebrates [
29]. Therefore, the expression pattern of this gene during zebrafish sex determination and differentiation is important. In non-mammalian vertebrates like birds and turtles,
sox9 is expressed in testes and down-regulated in ovaries during gonadal differentiation similar to the expression pattern seen in mammals [
30,
31]. In the present study, expression of
sox9a is found in several tissues including spleen, muscle, gall bladder, brain and eye of both male and female zebrafish. However,
sox9a is also expressed in gonad and gill in a male-specific manner. This is in accordance with a previous study in adult zebrafish showing different expression patterns for
sox9a and
sox9b in zebrafish gonads suggesting that
sox9a retained its function in the testis while
sox9b possibly acquired a different function in zebrafish ovary during evolution [
16]. The expression pattern of
sox9a in zebrafish during sex determination and differentiation segregated the fish into two groups. In fish of the high expresser group three peaks were observed. The first small peak seen at 4 dph is early in development and might be related to skeletal development, as
sox9 in mammals is also involved in this process [
16], but it could also be related to sex determination and differentiation. The peaks at 18 and 22 dph coincide with the expected time of gonadal differentiation, i.e. ovary to testis transformation in genetic males, which is a key event in zebrafish sex differentiation. In the mammalian bipotential gonad,
Sry initiates
sox9 expression and translocation from cytoplasm to nucleus which induces expression of
amh that is an important testis determining factor [
32]. The expression pattern of
sox9a seen in this study could indicate involvement of
sox9a in sex differentiation in zebrafish. However, no available data including those of the present study firmly indicates that
sox9a is the sex determining gene in zebrafish, but merely suggests that it might be involved in the sex signalling cascade and gonadal sex differentiation. This is in agreement with a study in medaka where the level of testicular type
sox9 (
sox9a2) expression in somatic cells is equally high in both sexes at the time when
dmy expression is initiated during early gonadal differentiation. However, during the period from 10–30 dph,
sox9 expression continues only in the Sertoli cells in male gonads, with a marked reduction in the XX gonads [
29].
In a study investigating
amh,
sox9a,
sox9b and
cyp19a1a expression in undifferentiated gonads of zebrafish, expression of all genes could be detected [
33], however, in the differentiated gonads a sexually dimorphic expression pattern was found;
sox9a and
amh were expressed in testis whereas
cyp19a1a was not. In ovaries,
sox9b and
cyp19a1a were expressed while
amh was not [
33]. Based on the expression pattern of these genes during sex differentiation, the authors suggested that 17 dph represents a transitional stage in zebrafish gonad development and by 31 dph gonads have differentiated into testes or ovaries [
33]. This is in accordance with the present study in which distinct peaks in both
cyp19a1a and
sox9a expression was found at 18 dph. Furthermore, these results are in agreement with those of a recent study of the molecular mechanism of the ovary-to-testes transition which indicated that all males go through the juvenile ovary phase until approximately 21 dpf (corresponding to 19 dph). However, they differ in the extent of commitment toward femaleness during this period and can be divided into three types of males based on the intensity, onset and duration of gonadal transformation [
3].
In a number of species including humans, mice, chickens, alligators and turtles, the
dmrt1 expression is limited to the gonads, and the expression is considerably up-regulated in the developing testes compared to ovaries. The timing of this up-regulation varies between species, but generally occurs in the late sex determining or early testis-differentiation period. This characteristic
dmrt1 expression pattern seen in different species indicates that this gene is specifically involved in the early formation of testes [
34]. The results from the present study correspond well with this general notion. We found peaks in expression of
dmrt1 in the high expresser group prior to the expected time of ovary to testis transformation. The expression peaks of
dmrt1 early in the sex determination and differentiation period could indicate that it is involved upstream in the signalling cascade that initiates sex determination and differentiation in zebrafish. In a recent model of medaka gonadal development it is shown that the
dmrt1 expression is detected from 10 dph and is thereby the first gene to be differentially expressed between males and females except for
DMY [
35]. This is in agreement with the results from the present study where we see a peak in
Dmrt1a expression at 10 dph. However, the idea that the
dmrt1 gene might be the sex-determining gene in non-mammalian vertebrates has been rejected based on the lack of
DMY (
Dmrt1bY) in other fish species than
Oryzias latipes, including
Oryzias celebensis,
Oryzias mekongenesis, guppy
Poecilia reticulata, pufferfish
Takifugu rubripes and zebrafish
Danio rerio [
34,
36]. The
dmrt1 gene most likely has a role downstream the sex determination event and it might be involved in testis development in teleost fish analogous to its putative role in mammalian species [
34]. This is in accordance with results from medaka where results indicated that
dmrt1 regulates spermatogonial differentiation [
37]. Sexual development in sex-reversed medaka gonads indicates that
DMY is not necessary for gonadal differentiation and that expression of
fig α and
dmrt1 correlate with the phenotypic differentiation of the gonads [
7]. The expression pattern of
fig α in the present study resembles to some extent that in the hermaphrodite fish
Kryptolebias marmoratus, where
fig α expression was very low until 39 dpf where expression peaks and remains high until 103 dpf which was the last day measured [
38]. Furthermore,
fig α showed no expression in males during gonadogenesis in mice [
39] which might be in accordance with the results in this study where we see low expression (close to the detection level) in the low expresser group. In mice,
fig α has been suggested to play a key role in preserving oocytes and normal ovarian development [
39] and in medaka
fig α is an oocyte specific marker [
38].
In this study, the pattern of
ar expression in zebrafish during sex determination and differentiation clearly segregated the fish into two groups. This is in accordance with a study in sea bass where differences in
ar expression were first encountered at 150 dph and became especially marked at 250 dph (corresponding with the time of sex differentiation) with higher expression in a male dominant group compared to a female dominant group (male and female dominant groups were based on consecutive size grading) [
22]. In the current study, we find peaks in ar expression in the high expresser group at 16 and 25 dph which is shortly after a peak in expression of
dmrt1 is observed and shortly before a peak in
sox9a expression. These peaks in expression of the different genes around 10–18 dph coincide with the expected initiation of transformation from ovary to testis in individuals developing into males. The difference in expression profiles between the two groups at the time of sex differentiation is most likely sex-related and corresponds to differences between males and females, suggesting an important role for
ar in sex differentiation. In a recent study of zebrafish gonad development, Hossain et al (2007) found
ar mRNA levels at 4 weeks post fertilization (their first measurement) to be similar in males and females. However, from 5–7 wpf the
ar expression was higher in the transforming testis [
24]. Furthermore, expression levels of
ar mRNA at different developmental stages has been studied in hermaphrodite fish species such as the protogynous wrasse,
Halichoeres trimaculatus [
40] and the protandrous black porgy,
Acanthopagrus schlegeli [
41]. In the black porgy, both the testicular tissue of the bisexual gonad and the functional testis exhibited higher
ar mRNA levels than the ovarian tissue. Conversely,
ar mRNA levels were lower in functional ovaries than in the ovarian part of the bisexual gonad, suggesting association of the decrease in
ar mRNA levels with protandrous sex change in this fish species [
41]. The ubiquitous expression of
ar in adult zebrafish of both sexes found in this study is in accordance with [
24] and we found
ar expression in all investigated tissue of adult zebrafish (gonad, brain, liver, kidney, skin, muscle, eye) of both males and females.
Also, the ovarian form of aromatase,
cyp19a1a was suggested to be involved in sex determination and differentiation [
33].
cyp19a1a expression was found in undifferentiated gonads at 17 dpf which is expected based on the juvenile ovarian gonads. However, at 31 dpf the differentiated ovary showed specific expression of
cyp19a1a in cells surrounding the oocytes [
33]. Furthermore, at 31 dpf the differentiated testis showed no expression of
cyp19a1a [
33]. The results presented in this study correspond well with the previous studies of
cyp19a1a expression. We find two peaks in expression during the sex determination and differentiation period (at 18 and 30 dph) in the high expresser group. Furthermore, we find that the expression of
cyp19a1a in both the high and low expresser group is very low (almost undetectable in the low expresser group) at the time of gonadal transformation at approximately 19–21 dph. In accordance, a recent study showed
cyp19a1a expression around oocytes in juvenile ovary, this expression was down-regulated and could no longer be detected when gonadal transformation was initiated [
42].
The inter-individual grouping of the expression of
cyp19a1b during sex determination and differentiation into high and low expressers observed in the present study is in accordance with previous studies [
1,
27,
43]. In the present study we see one distinct peak in
cyp19a1b expression at 14 dph for the high expresser group. This does not correspond completely with the peaks found in the high expresser group during sex determination and differentiation in the study by Trant et al. (2001), where
cyp19a1b peaks were seen at 4–5 dpf, 13 dpf (small peak) and 25 dpf and from 34 – 40 dpf increase (with 2.5 dpf = 0 dph) [
1]. The small peak at 13 dpf is most likely corresponding to the peak we see at 14 dph. However, the general expression pattern is not similar in the two studies. The finding that individual
cyp19a1b expression during sex determination and differentiation distributes into two groups indicates that the expression of
cyp19a1b (the brain form) might be associated with sexual differentiation in zebrafish as previously suggested by Trant et al. (2001). In contrast, Gato-Kazeto et al. (2004) investigated high and low expressers of
cyp19a2 (the brain form) and found that it did not correlate with sex in adult zebrafish [
27] Furthermore, a recent study on
cyp19a1b expression in brain of adult zebrafish found that the expression levels were in the same range in males and females [
44]. This is in accordance with the present study that indicates high brain expression of
cyp19a1b in adults of both sexes. Due to the conflicting results regarding the involvement of
cyp19a1b in gonadal sex differentiation,
cyp19a1b is not included in the calculation of ratios between expected male and female genes in this study.
The coinciding low expression of
sox9a,
dmrt1,
ar,
fig α and
cyp19a1a (in the high expresser groups) at 19–20 dph is just prior to the time of oocyte apoptosis in individuals developing into males. This could indicate that genes involved in the sex differentiation might be down-regulated at this time in development. Interestingly, it has previously been shown that the
vas gene (which is a germ cell marker) was also down-regulated at 23 dpf [
45]. Furthermore, the genes expected to be highly expressed in females (
fig α and
cyp19a1a) both had increased expression levels from 22 dph. Also, two of the three genes expected to be highly expressed in males (
ar and
sox9a) have a peak in expression at 22 dph. The last gene expected to be highly expressed in males (
dmrt1) does not have a significant peak in expression after the general down regulation at 19–20 dph; this could indicate that dmrt1 (which shows two significant peaks in expression before 19–20 dph, at 10 and 14 dph) is involved in gonadal sex differentiation upstream in the signalling cascade.