- Email: [email protected]
DMRT1 expression during gonadal differentiation and spermatogenesis in the rainbow trout, Oncorhynchus mykiss
Biochimica et Biophysica Acta 1493 (2000) 180^187
www.elsevier.com/locate/bba
DMRT1 expression during gonadal di¡erentiation and spermatogenesis in the rainbow trout, Oncorhynchus mykiss Oriane Marchand a , Marina Govoroun b , Helena D'Cotta c , Oonagh McMeel d , Jean-Jacques Lareyre b , Alain Bernot e , Vincent Laudet a , Yann Guiguen b; * a
d
UMR 5665 du CNRS, Ecole Normale Supe¨rieure de Lyon, 46, Alle¨e d'Italie, 69364 Lyon Cedex 07, France b Station INRA SCRIBE, Campus de Beaulieu, 35042 Rennes Cedex, France c CIRAD-EMVT, Campus de Beaulieu, 35042 Rennes Cedex, France NDC (National Diagnostic Centre), National University of Ireland, Galway, University Road, Galway, Ireland e Centre National de Se¨quenc° age, 2 rue gaston Cre©mieux, B.P. 191, 91006 Evry Cedex, France Received 23 December 1999; received in revised form 29 June 2000 ; accepted 17 July 2000
Abstract DMRT1 has been suggested to be the first conserved gene involved in sex differentiation found from invertebrates to human. To gain insight on its implication for fish gonadal differentiation, we cloned a DMRT1 homologue in the rainbow trout, Oncorhynchus mykiss (rtDMRT1), and showed that this gene is expressed during testicular differentiation, but not during ovarian differentiation. After 10 days of steroid treatment, expression was shown to be decreased in estrogen-treated male differentiating gonads but not to be restored in androgentreated differentiating female gonads. This clearly reinforces the hypothesis of an important implication for DMRT1 in testicular differentiation in all vertebrates. In the adults a single 1.5 kb transcript was detected by Northern blot analysis in the testis, and its expression was found to be sustained throughout spermatogenesis and declined at the end of spermatogenesis (stage VI). Along with this expression in the testis we also detected by reverse transcriptase^polymerase chain reaction a slight expression in the ovary. We also obtained new DM-domain homologous sequences in fish, and their analysis suggest that at least four different genes bearing `DM-domain' (DMRT genes) exist in fish just as in all vertebrate genomes. ß 2000 Elsevier Science B.V. All rights reserved. Keywords : Sex di¡erentiation; DMRT1 ; DM domain; Spermatogenesis; Rainbow trout; Oncorhynchus mykiss
1. Introduction DMRT1 is a gene encoding a protein with a DNA-binding motif called DM domain [1]. This gene has been recently characterized in human as the counterpart of two genes involved in invertebrate sex di¡erentiation namely, doublesex (dsx) in Drosophila melanogaster [2] and mab-3 in Caenorhabditis elegans [3]. Based on this evolutionary conservation, testis speci¢c expression [1] and its chromosomal localization on a region associated with XY sex reversal in human [4^6] and on the sex chromosome Z in chicken [7], DMRT1 has been suggested to be involved in sex di¡erentiation. Recently, its expression in the di¡erentiating testis has also been shown to be consistent with a role in testicular di¡erentiation in di¡erent vertebrates spe-
* Corresponding author. Fax: +33-2-2348-5020; E-mail : [email protected]
cies including some mammals [8^11], birds [8,9], and reptiles [8,12]. In some ¢sh, including rainbow trout (Oncorhynchus mykiss), the availability of genetic all-male and all-female populations through the establishment of new viable and fertile males with YY or XX genotypes [13] o¡ers a unique possibility for studying sex di¡erentiation. These genetic mono-sex populations enable studies on a large number of mono-sex ¢sh before the onset of the ¢rst histological features of gonadal di¡erentiation. Furthermore, ¢sh gonadal di¡erentiation is very sensitive to steroids [14] and the phenotypic sex can be manipulated easily by hormonal treatments [15]. These advantages, and the fact that no DMRT1 gene expression has been reported in ¢sh, led us to explore DMRT1 gene expression during natural and steroid-induced sex di¡erentiation in the rainbow trout. Furthermore, during the course of this study we also found some other type of DM-domain sequences in rainbow trout, and since trout is known to have multiple gene families due to a more recent tetraploidization event
0167-4781 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 7 8 1 ( 0 0 ) 0 0 1 8 6 - X
BBAEXP 93447 28-8-00
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
[16], this led us to analyze these new sequences along with other vertebrate DM-domain sequences. 2. Materials and methods 2.1. Animals All-female and all-male populations of rainbow trout were obtained by means of fertilizing normal eggs with sperm of phenotypic XX or YY males. The YY male genotype is viable in rainbow trout [13] and these XX and YY males were produced as previously described [17]. Steroid treatments were carried out at the INRA experimental ¢sh farm (Drennec, France), by dietary administration starting from the onset of the ¢rst feeding, i.e., reabsorption of the vitellin vesicle at 55 days post-fertilization (dpf). 11L-Hydroxyandrostenedione (11LOHv4, 10 mg/kg of food) and 17L-estradiol (E2, 20 mg/kg of food) (Sigma) were added to the food (dry pellet food, Aqualim) in ethanolic solution with 40 ml of ethanolic solution per kg of food. For tissue sampling, ¢sh were killed either by decapitation or by a prolonged overdosed anesthesia in 0.5% 2phenoxy-ethanol. For histological analysis gonads were ¢xed in Bouin-Holland £uid, dehydrated in aqueous ethanol 70% and 95%, then in ethanol/butanol (5:95) and ¢nally butanol. Tissues were embedded in para¤n, and 5Wm sections were stained with Regaud hematoxylin [18]. For the semi-quantitative reverse transcriptase^polymerase chain reaction (RT^PCR) assay in rainbow trout, 200^400 gonads were dissected out under the dissecting microscope at weekly intervals from 55 to 127 days post-fertilization (dpf). Gonads were immediately frozen in liquid nitrogen and stored at 380³C until RNA extraction. For the analysis of natural gonadal di¡erentiation, gonads were sampled in all-male (1830 gonads) and all-female (1860 gonads) populations at around 55 dpf ( þ 5 dpf). For the study of the androgen- and estrogen-induced di¡erentiation, gonads were sampled at 65 dpf ( þ 2 dpf) in all-female control populations (785 and 755 gonads), allmale control populations (752 and 781 gonads), all-female populations treated with 11LOHv4 (10 mg/kg of food) for 10 days from the onset of the ¢rst feeding (905 and 845 gonads) and all-male populations treated with E2 (20 mg/ kg of food) for 10 days from the onset of the ¢rst feeding (778 and 500 gonads). Stages of rainbow trout spermatogenesis were determined by histological analysis according to Billard and Esca¡re [19]. Stage I corresponds to immature ¢sh, stages II and III to early spermatogenesis, stages IV and V to mid-spermatogenesis and stage VI to spermiation. 2.2. cDNA cloning Degenerate primers were designed (underlined with dot-
181
ted lines in Fig. 2) according to a nucleotide alignment of di¡erent DM domains including those of D. melanogaster dsx, C. elegans mab-3, and human DMRT1 (sense primer sequences, DMRT51: TCSCCIMGSMYRCCSAAITGYGC and DMRT52: GCCTCRCCGYTIAARGGICACAA; antisense primer sequence, DMRT31: CYMARRGCIACYTGIGCIGCCAT). Touchdown PCR was carried out as described [20], using cDNAs reverse transcribed from mRNA (1 Wg) originating from various rainbow trout tissues (testis, ovary, brain, heart, muscle, liver and kidney) and genomic DNA. PCR products of the expected size were then directly subcloned into PCR2.1 vector (InVitrogen). A PCR fragment, homologous to DMRT1, was then used as a probe to screen a VZAP rainbow trout testis cDNA library according to standard methods [21], and the positive clones were ¢nally sequenced on both strands. 2.3. Sequences analysis Phylogenetic reconstruction was made with PHYLO_ Win [22] using the Neighbor-Joining Method [23] on amino acid sequences using Poisson correction and a global gap removal option. Two other methods, namely Maximum Parsimony (also implemented in PHYLO_Win) and a Quartet Maximum Likelihood [24], were also used. Bootstrap analysis with 1000 replicates was used to assess strength of nodes in the tree [25]. Accession numbers of the sequences are as follows : Gallus DMRT1, AF123456; Oncorhynchus DMRT1, AF209095; Danio DMRT2 (TERRA), AF080622; Oncorhynchus DMRT2, AF209096; Oncorhynchus DMRT4, AF209097; Tetraodon DMRT4, AJ251455; Human DMRT3, AF193872; Tetraodon DMRT3, AJ251456 ; Drosophila DSX, M25292; Bactrocera DSX, AF029676; Caenorhabditis MAB-3, U49831. The Homo sapiens chromosome 1 sequence containing a human DMRT4 homologue is referenced AL049637. 2.4. Semi-quantitative RT^PCR analysis Semi-quantitative RT^PCR was performed as previously described [17]. Brie£y, Poly(A) RNA were isolated directly from frozen gonads using Dynabeads (Dynal, Oslo, Norway) as described by the manufacturer. After reverse transcription, standard PCR reactions including negative (water only) and genomic DNA controls (with a low optimized number of cycles, i.e., 20 cycles for L-actin and 15 for rtDMRT1) were performed with speci¢c primers for each gene of interest i.e., rtL-actin (rtL-actinsense, 5P-AAAGACCCTGAGTTCATCATGC-3P; rtLactin-antisense, 5P-CCCAGTCTCCACTAATCCCA-3P), rtGAPDH (rtGAPDH-sense: 5P-GACCTCTGTGTTGGAATCAACG-3P; rtGAPDH-antisense, 5P-TCCCATTGCAGCTTTCTGGAGC-3P), and rtDMRT1 (underlined with double lines in Fig. 2, rtDMRT1-sense, 5P-TTCCTG-
BBAEXP 93447 28-8-00
182
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
Fig. 1. (A) Phylogenetic tree connecting some `DM domain' sequences and Clustal alignment of these DM-domain sequences. Numbers in each branch represent the bootstrap values obtained by Neighbor-Joining, Maximum Parsimony and Maximum Likelihood, respectively. Branch lengths are proportional to evolutionary distances. (B) Amino acid sequences are compared to human DMRT1 gene. Dots indicate identical residues and gaps are represented by dashes. The sequence of Caenorhabditis elegans C34D1.1 related to MAB-3 gene is not represented but is used as a root for the construction of the phylogenic tree. C. elegans MAB-3b corresponds to the second DM domain of the MAB-3 gene.
CAACCGGCCAGAAG-3P; rtDMRT1-antisense, 5P-AACAACCTCCTGACTGGACA-3P). One Wl of each PCR reaction was dotted on nylon membrane and hybridized with the corresponding [K-32 P]dCTP-labeled probe. Membranes were then quantitatively analyzed using an instant imager (Packard) and data were expressed as logarithms of [rtDMRT1]/[L-actin] ratios. In our standard PCR conditions (94³C for 1 min, 58³C for 1 min, 72³C for 30 s), PCR reactions always ampli¢ed a single band of the expected size and no ampli¢cation product
was detected either in our water or genomic DNA controls. 2.5. Northern and virtual northern analysis Northern blots were essentially performed as described [21], except that hybridization was performed in ULTRAhyb solution (Ambion). Due to the minute amount of tissue and hence RNA obtained from the 1800 di¡erentiating gonads, techniques such as a standard Northern blot
BBAEXP 93447 28-8-00
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
183
Fig. 2. Nucleotide sequence of rainbow trout, Oncorhynchus mykiss, rtDMRT1, with the corresponding amino acids translation represented in bold type. Start and stop codon are in bold italics and the DM domain, as previously described in [5], is underlined. The sequences underlined with dots correspond to the amino acids used for the design of the degenerate primers. The sequences underlined with double lines correspond to rtDMRT1 sense and antisense primers used for semi-quantitative RT^PCR.
and RNase protection assays were not possible because they were not sensitive enough. Thus more sensitive techniques should be applied such as virtual Northern. This technique has been proved to be absolutely feasible and sound for quanti¢cation purposes [26,27]. Virtual Northern was carried out according to the manufacturer (SMART PCR cDNA Synthesis kit, Clontech) and as previously described [17], using 1 Wg of total RNA. cDNA were loaded onto 1% TBE/agarose gels and transferred to nylon membranes (roughly from 200 to 600 ng for Fig. 3 and between 1 and 2 Wg for Fig. 4). The membrane was then hybridized in ULTRAhyb1 solution with
[K-32 P]dCTP-labeled probes. The DMRT1 probe consisted of the entire coding sequence of rtDMRT1 and reprobing of the membrane with glyceraldehyde-3-phosphate dehydrogenase cDNA (rtGAPDH, accession number AF027130) or L-actin was used as a control. 3. Results 3.1. DM-domain sequences The cloning of a DMRT1 homologous cDNA in rain-
BBAEXP 93447 28-8-00
184
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
Fig. 3. Expression of rtDMRT1 during gonadal di¡erentiation of rainbow trout. (a) Virtual Northern blot analysis of rtDMRT1 and rtGAPDH gene expression in rainbow trout di¡erentiating gonads. Increasing amounts of cDNA were loaded onto the gel (4, 6, 8, 10 and 12 Wl of the PCR reaction, at a concentration around 50 ng/Wl). (b) Histological cross-section of the gonads sampled at 60 dpf showing an absence of meiotic oocytes. Gc, gonia cells; Ic, interstitial cells; Bc, blood cells; Mes, mesogonium. (c) RT^PCR semi-quantitative analysis of rtDMRT1 (represented as mean þ S.D.) of the ratio over L-actin with an arbitrary logarithmic scale), before (pre-di¡erentiation), during (di¡erentiation), and after (post-di¡erentiation) histological differentiation. M, male ; F, female. N represents the number of samples analyzed.
bow trout was performed using RT^PCR and degenerate primers designed against conserved regions of di¡erent DM domains. Three types of sequence showing high levels of sequence identity with di¡erent DM domains were found in rainbow trout (in testis, ovary and genomic DNA). The relationships existing between these various sequences (Fig. 1) indicated that one of them (rtDMRT1 found only in testis, primers DMRT52UDMRT31) clearly clustered with the human and chicken DMRT1 cDNA sequences. A second DM homologue domain (found in ovary, primers DMRT51UDMRT31) is clearly the homologue of the zebra¢sh, Danio rerio, TERRA [28] and human DMRT2 [5,28] genes (rtDMRT2). Finally, the third type of DM domain homologous sequences (found both in ovary and genomic DNA, primers DMRT51UDMRT31) is related to a human gene found in the sequence of chromosome 1, that we tentatively named DMRT4 (accession number AL049637). We also isolated two sequences from another ¢sh species, the pu¡er¢sh, T. nigroviridis, and found that these additional sequences (pDMRT4 and pDMRT3) are related to human and rainbow trout DMRT4 and also to the recently described human DMRT3 sequences.
3.2. rtDMRT1 sequence In order to look at the implication for DMRT1 in gonadal di¡erentiation, we used the rtDMRT1 PCR fragment to screen a rainbow trout testis cDNA library. Among a few positives obtained, we sequenced a 1138 bp clone (Fig. 2) encoding a predicted 325-amino-acid protein that exhibits 90% and 87% sequence identity, respectively, with chicken and human DMRT1, but this identity is only found inside the DM domain. Outside this domain, no important sequence similarity was detected and especially the 3P end proline/serine rich region described in other species was not found in the rainbow trout sequence. 3.3. rtDMRT1 expression during gonadal di¡erentiation Semi-quantitative RT^PCR and virtual Northern analysis were carried out on genetic mono-sex populations to analyze rtDMRT1 gene expression during gonadal di¡erentiation. rtDMRT1 was shown to be highly expressed in the di¡erentiating testis when compared to the di¡erentiating ovary (Fig. 3a,c). Virtual Northern hybridization
BBAEXP 93447 28-8-00
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
with rtDMRT1 probe revealed one major band around 1.5 kb and two faint additional signals in the di¡erentiating testis (represented by arrows in Fig. 3a). Only the upper two bands were expressed in female di¡erentiating gonads but at an extremely low level. Histology of the gonads of the all-male and all-female populations sampled for the virtual Northern analysis is presented in Fig. 3b. This analysis con¢rmed that these gonads were still histologically undi¡erentiated at that time, i.e., without any sign of meiotic oocyte (as the early detection of oocyte meiosis in females is the main histological criteria used for assessing gonadal di¡erentiation in ¢sh [30,31]). By semi-quantitative RT^PCR, we also con¢rmed this early testicular expression and shown that this expression was sustained until the ¢rst male meiosis around 120 dpf (Fig. 3c).
185
Fig. 5. Expression of rtDMRT1 (top panel) in some adult tissues of rainbow trout by Northern blot. Two Wg of mRNA was loaded onto each lane from the following tissues: T, testis ; O, ovary; S, skin; H, heart; L, liver ; pK, posterior kidney; aK, anterior kidney. Reprobing with L-actin (bottom panel) was used as a loading control.
Virtual Northern analysis (Fig. 4) was also carried out on di¡erentiating gonads sampled from genetic mono-sex populations treated either with estrogen (E2, 20 mg/kg of food) or androgens (11LOHv4, 10 mg/kg of food). In the estrogen-treated all-male population, rtDMRT1 was found after 10 days of steroid treatment to be decreased 2.5-fold when compared to the all-male untreated population. Conversely, expression was not detectable, at short exposure time, in both androgen-treated and untreated control all-female populations. After an over exposition of this virtual Northern blot, a slight rtDMRT1 expression was detected but without any important di¡erence between treated and untreated female populations. 3.4. rtDMRT1 expression in some adult tissues In adult, rtDMRT1 expression was detected by Northern blot in the testis, with a single 1.5 kb transcript (Fig. 5), but was not detected in the other tissues tested (ovary, heart, intestine, skin, kidney and liver). Using RT^PCR analysis (Fig. 6), we con¢rmed an important expression in the testis but we also detected a slight expression in the previtellogenetic ovary. No expression was detected
Fig. 4. Expression of rtDMRT1 during steroid-induced gonadal di¡erentiation in rainbow trout by virtual Northern blot analysis (each lane contains about 1^2 Wg of cDNA). Reprobing with L-actin (bottom panel) was used as a loading control. (a) Expression of rtDMRT1 following a feminizing estrogen treatment (17L-estradiol, 20 mg/kg of food) on an all-male (M), ME2: estrogen-treated male population. (b) Expression of rtDMRT1 following a masculinizing androgen treatment (11L-hydroxyandrostenedione, 10 mg/kg of food) on an all-female (F), F.11Lv4: androgen-treated female population. The right panel is an over-exposition of the virtual Northern shown in the left panel.
Fig. 6. (Top panels) Expression of rtDMRT1 in some adult tissues of rainbow trout by RT^PCR. RT^PCR was carried out with 25 or 30 cycles and the 25-cycle experiment was transferred on nylon membrane and hybridized with an rtDMRT1 probe. T, testis; O, ovary; B, brain; H, heart; L, liver ; M, muscle; aK, anterior kidney; pK, posterior kidney. RT^PCR also included negative (3) and genomic DNA (G) controls. RT^PCR with rtGAPDH primers (bottom panel) was used as a control.
BBAEXP 93447 28-8-00
186
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
4.2. rtDMRT1 expression.
Fig. 7. (Top) Expression of rtDMRT1 during rainbow trout spermatogenesis by Northern blot analysis. About 20 Wg of total RNA was loaded onto each lane. Reprobing with RNA 28S was used as a loading control. The ratio of rtDMRT1 over RNA 28S (bottom) was estimated following quanti¢cation of rtDMRT1 and RNA 28S signal using an instant imager.
by RT^PCR for the following tissues: brain, heart, liver, muscle, and kidney (Fig. 6). Expression analysis by Northern blot during spermatogenesis (Fig. 7) revealed that rtDMRT1 is highly expressed at comparable levels in immature testis (stage I), early spermatogenesis (stages II and III) and mid spermatogenesis (stages IV and V). However, during spermiation (stage VI) this expression is strongly reduced (to less than 10% of the average of other stages). 4. Discussion 4.1. DM-domain gene family We have isolated a total of four di¡erent DM-domain genes in ¢sh. Their homology within their DM domain clearly identi¢es them as members of the DMRT family (DMRT stands for Doublesex and MAB-3 Related Transcription factor according to the proposed terminology [5]). The phylogenetic analysis reveals that these sequences belong to four di¡erent genes. The full-length sequence from trout described in this paper is clearly a homologue of the human and chicken DMRT1. These four DMRT genes are found in at least mammals and ¢shes (e.g., [1,5,7,28]), and thus it can be inferred that they result from an ancient duplication that arose very early in vertebrate evolution as has been found for other gene families [32]. Considering the chromosomal location of these genes in human, the most divergent gene in our analysis, DMRT3, is the only one to be located on human chromosome 1, whereas the three others are clustered in chromosome 9 [6]. This may suggest that they arose by tandem duplication, a view that is not in agreement with the hypothesis of four paralogous genes common to all vertebrates.
DMRT1 is now well known as a transcription factor likely to be involved in sex di¡erentiation in many vertebrates [7^12]. Our results clearly demonstrate, for the ¢rst time in ¢sh, that rainbow trout DMRT1 (rtDMRT1) is also highly expressed during testicular di¡erentiation but not during ovarian di¡erentiation. This testicular expression is detected as early as 2^3 weeks before the ¢rst histological criteria of sex di¡erentiation in ¢sh that is the early oocyte meiosis detection in di¡erentiating ovaries [30,31]. Expression during steroid-induced di¡erentiation demonstrate that a short-term feminizing treatment with estradiol-17L is able to downregulate the expression of rtDMRT1. This downregulation by a feminizing agent is in agreement with the idea that DMRT1 is really a testicular di¡erentiation gene and, in a similar manner, masculinizing temperature has been found to increase DMRT1 expression in a temperature-dependent sex determination (TSD) species of turtle [12]. However, in rainbow trout this downregulation by estrogens was found to be only partial and we did not ¢nd any stimulation of rtDMRT1 expression following masculinization of a genetic female population with androgens. This may likely be due to our sampling date that is only 10 days after the beginning of the treatment. Such short-term e¡ects of steroid treatments on some gene transcriptions have been already shown during gonadal di¡erentiation in rainbow trout (Govorun et al., unpublished results) and in the case of rtDMRT1 this absence of important e¡ects may indicate an indirect action of these steroids on the regulation of DMRT1 gene expression. This could suggest that DMRT1 is not a direct downstream gene under the control of steroids, and may have a more upstream position in the sex di¡erentiation cascade. Upstream genes involved in that cascade are generally thought to be less conserved genes [33], as is the case for instance for SRY in mammals, where only the HMG box shows a good conservation [34]. However, the non-vertebrate homologues of DMRT1, dsx in Drosophila and MAB-3 in nematodes, are thought to be more downstream genes in the cascade of sex di¡erentiation [35]. The exact position of DMRT1 in the sex di¡erentiation cascade of vertebrates remains to be elucidated, but its expression in normal testis development [8^11], and at the same time as Sry, i.e., just before testicular di¡erentiation [8], would favor an upstream position and a major role for that gene. The hypothesis of DMRT1 activating Sry has even been proposed [9], based on an earlier detection of DMRT1 gene expression in mouse (E9.5 for DMRT1 and E10.5 for Sry). However, in ¢sh, as in other non-mammalian vertebrates, no Sry equivalent has been found, and thus DMRT1 may even play a more crucial role in these species than in mammals. In rainbow trout adult tissues, the expression of rtDMRT1 was found in the testis as previously found in human [1] and mouse [9], with a transcript size of 1.5 kb,
BBAEXP 93447 28-8-00
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
identical to the chicken transcript [7] but smaller that the human and mouse ones [1,10]. However, in rainbow trout there is also a small expression of rtDMRT1 in the ovary. This expression in the developing ovary is 80 times less than in the developing testis before histological di¡erentiation but this di¡erence is reduced to a factor of 40 during, and to a factor of 10 after histological sex di¡erentiation. This was also con¢rmed by RT^PCR analysis in adult tissues. Expression in developing ovaries has also been found in the chicken and alligator, although a higher expression was always detected in developing male embryos [8,9]. In rainbow trout, rtDMRT1 is highly expressed during spermatogenesis until late spermatogenesis, and decreases thereafter at spermiation. In mouse, expression of DMRT1 is localized both in Sertoli cells and primordial germ cells [9] but there is no indication as to whether expression is also sustained in germ cells at later stages such as spermatid or spermatozoa. The decrease in rtDMRT1 expression seen in rainbow trout at spermiation may thus either indicate a dilution of an important rtDMRT1 expression in Sertoli cells that are considerably diluted by germinal cells during the late stages of spermatogenesis, or a sudden decrease in the rtDMRT1 expression in spermatozoon that are the major cell type at spermiation. 4.3. Conclusion These results clearly demonstrate a sexually dimorphic expression of rtDMRT1 during gonadal di¡erentiation in rainbow trout, and this reinforces the hypothesis of an important implication for DMRT1 in testicular di¡erentiation in all vertebrates [1,5^12]. However, the occurrence of multiple genes with `DM domains' in ¢sh would also require further study since like in mammals [1,5,6,29], at least four `DM domain'-related genes exist in ¢sh, and these may also have implications for developmental processes. Acknowledgements This work was supported by INRA, and by a European Community project, PL 97-3796 (Y.G.), as well as by CNRS and ENS (V.L.). We thank Jean-Marc Vanacker, Franck Delaunay and Alexis Fostier for critical reading of the manuscript. References [1] C.S. Raymond, C.E. Shamu, M.M. Shen, K.J. Seifert, B. Hirsch, J. Hodgkin, D. Zarkower, Nature 391 (1998) 691^695.
187
[2] K.C. Burtis, B.S. Baker, Cell 56 (1989) 997^1010. [3] M.M. Shen, J. Hodgkin, Cell 54 (1988) 1019^1031. [4] R. Veitia, M. Nunes, R. Brauner, M. Doco-Fenzy, O. Joanny-Flinois, F. Jaubert, S. Lortat-Jacob, M. Fellous, K. McElreavey, Genomics 41 (1997) 271^274. [5] C.S. Raymond, E.D. Parker, J.R. Kettlewell, L.G. Brown, D.C. Page, K. Kusz, J. Jaruzelska, Y. Reinberg, W.L. Flejter, V.J. Bardwell, B. Hirsch, D. Zarkower, Hum. Mol. Genet. 8 (1999) 989^996. [6] C. Ottolenghi, R. Veitia, L. Quintana-Murci, D. Torchard, L. Scapoli, N. Souleyreau-Therville, J. Beckmann, M. Fellous, K. McElreavey, Genomics 64 (2000) 170^178. [7] I. Nanda, Z. Shan, M. Schartl, D.W. Burt, M. Koehler, H. Nothwang, F. Grutzner, I.R. Paton, D. Windsor, I. Dunn, W. Engel, P. Staeheli, S. Mizuno, T. Haaf, M. Schmid, Nat. Genet. 21 (1999) 258^259. [8] C.A. Smith, P.J. McClive, P.S. Western, K.J. Reed, A.H. Sinclair, Nature 402 (1999) 601^602. [9] C.S. Raymond, J.R. Kettlewell, B. Hirsch, V.J. Bardwell, D. Zarkower, Dev. Biol. 215 (1999) 208^220. [10] A. De Grandi, V. Calvari, V. Bertini, A. Bulfone, G. Peverali, G. Camerino, G. Borsani, S. Guioli, Mech. Dev. 90 (2000) 323^326. [11] B. Moniot, P. Berta, G. Scherer, P. Sudbeck, F. Poulat, Mech. Dev. 91 (2000) 323^325. [12] J.R. Kettlewell, C.S. Raymond, D. Zarkower, Genesis 26 (2000) 174^ 178. [13] B. Chevassus, A. Devaux, D. Chourrout, B. Jalabert, J. Hered. 79 (1988) 89^92. [14] T. Yamamoto, in: W.S. Hoar, D.J. Randall (Eds.), Fish Physiology, Academic Press, New York, 1969, pp. 117^175. [15] G.A. Hunter, E.M. Donaldson, in: W.S. Hoar, D.J. Randall (Eds.), Fish Physiology, vol. IX. Reproduction, Academic press, New York, 1983, pp. 223^303. [16] J. Wittbrodt, A. Meyer, M. Schartl, BioEssays 20 (1998) 511^515. [17] Y. Guiguen, J.F. Baroiller, M.J. Ricordel, K. Iseki, O.M. McMeel, S.A.M. Martin, A. Fostier, Mol. Reprod. Dev. 54 (1999) 154^162. [18] M. Gabe, Techniques Histologiques, Masson and Cie, Paris, 1968. [19] R. Billard, A.M. Escra¡re, Bull. Fr. Pisci. 256 (1975) 111^118. [20] H. Escriva, R. Sa¢, C. Hanni, M.C. Langlois, P. Saumitou-Laprade, D. Stehelin, A. Capron, R. Pierce, V. Laudet, Proc. Natl. Acad. Sci. USA 94 (1997) 6803^6808. [21] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning : A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989. [22] N. Galtier, M. Gouy, C. Gautier, Comput. Appl. Biosci. 12 (1989) 543^548. [23] N. Saitou, M. Nei, Mol. Biol. Evol. 4 (1987) 406^425. [24] K. Strimmer, A. Von Haeseler, Mol. Biol. Evol. 13 (1996) 964^969. [25] J. Felsenstein, Evolution 39 (1985) 783^791. [26] W.O. Endege, K.E. Steinmann, L.A. Boardman, S.N. Thibodeau, R. Schlegel, Biotechniques 26 (1999) 542^550. [27] O. Franz, I. Bruchhaus, T. Roeder, Nucleic Acids Res. 27 (1999) e3. [28] A. Meng, B. Moore, H. Tang, B. Yuan, S. Lin, Development 126 (1999) 1259^1268. [29] C. Ottolenghi, R. Veitia, M. Barbieri, M. Fellous, K. McElreavey, Genomics 64 (2000) 179^186. [30] J.F. Baroiller, Y. Guiguen, A. Fostier, Cell. Mol. Life Sci. 55 (1999) 910^931. [31] M. Nakamura, T. Kobayashi, X. Chang, Y. Nagahama, J. Exp. Zool. 281 (1998) 362^372. [32] V. Laudet, J. Mol. Endocrinol. 19 (1997) 207^226. [33] I. Marin, B.S. Baker, Science 281 (1998) 1990^1994. [34] P. Koopman, Cell. Mol. Life Sci. 55 (1999) 839^856. [35] W. Yi, D. Zarkower, Development 126 (1999) 873^881.
BBAEXP 93447 28-8-00
www.elsevier.com/locate/bba
DMRT1 expression during gonadal di¡erentiation and spermatogenesis in the rainbow trout, Oncorhynchus mykiss Oriane Marchand a , Marina Govoroun b , Helena D'Cotta c , Oonagh McMeel d , Jean-Jacques Lareyre b , Alain Bernot e , Vincent Laudet a , Yann Guiguen b; * a
d
UMR 5665 du CNRS, Ecole Normale Supe¨rieure de Lyon, 46, Alle¨e d'Italie, 69364 Lyon Cedex 07, France b Station INRA SCRIBE, Campus de Beaulieu, 35042 Rennes Cedex, France c CIRAD-EMVT, Campus de Beaulieu, 35042 Rennes Cedex, France NDC (National Diagnostic Centre), National University of Ireland, Galway, University Road, Galway, Ireland e Centre National de Se¨quenc° age, 2 rue gaston Cre©mieux, B.P. 191, 91006 Evry Cedex, France Received 23 December 1999; received in revised form 29 June 2000 ; accepted 17 July 2000
Abstract DMRT1 has been suggested to be the first conserved gene involved in sex differentiation found from invertebrates to human. To gain insight on its implication for fish gonadal differentiation, we cloned a DMRT1 homologue in the rainbow trout, Oncorhynchus mykiss (rtDMRT1), and showed that this gene is expressed during testicular differentiation, but not during ovarian differentiation. After 10 days of steroid treatment, expression was shown to be decreased in estrogen-treated male differentiating gonads but not to be restored in androgentreated differentiating female gonads. This clearly reinforces the hypothesis of an important implication for DMRT1 in testicular differentiation in all vertebrates. In the adults a single 1.5 kb transcript was detected by Northern blot analysis in the testis, and its expression was found to be sustained throughout spermatogenesis and declined at the end of spermatogenesis (stage VI). Along with this expression in the testis we also detected by reverse transcriptase^polymerase chain reaction a slight expression in the ovary. We also obtained new DM-domain homologous sequences in fish, and their analysis suggest that at least four different genes bearing `DM-domain' (DMRT genes) exist in fish just as in all vertebrate genomes. ß 2000 Elsevier Science B.V. All rights reserved. Keywords : Sex di¡erentiation; DMRT1 ; DM domain; Spermatogenesis; Rainbow trout; Oncorhynchus mykiss
1. Introduction DMRT1 is a gene encoding a protein with a DNA-binding motif called DM domain [1]. This gene has been recently characterized in human as the counterpart of two genes involved in invertebrate sex di¡erentiation namely, doublesex (dsx) in Drosophila melanogaster [2] and mab-3 in Caenorhabditis elegans [3]. Based on this evolutionary conservation, testis speci¢c expression [1] and its chromosomal localization on a region associated with XY sex reversal in human [4^6] and on the sex chromosome Z in chicken [7], DMRT1 has been suggested to be involved in sex di¡erentiation. Recently, its expression in the di¡erentiating testis has also been shown to be consistent with a role in testicular di¡erentiation in di¡erent vertebrates spe-
* Corresponding author. Fax: +33-2-2348-5020; E-mail : [email protected]
cies including some mammals [8^11], birds [8,9], and reptiles [8,12]. In some ¢sh, including rainbow trout (Oncorhynchus mykiss), the availability of genetic all-male and all-female populations through the establishment of new viable and fertile males with YY or XX genotypes [13] o¡ers a unique possibility for studying sex di¡erentiation. These genetic mono-sex populations enable studies on a large number of mono-sex ¢sh before the onset of the ¢rst histological features of gonadal di¡erentiation. Furthermore, ¢sh gonadal di¡erentiation is very sensitive to steroids [14] and the phenotypic sex can be manipulated easily by hormonal treatments [15]. These advantages, and the fact that no DMRT1 gene expression has been reported in ¢sh, led us to explore DMRT1 gene expression during natural and steroid-induced sex di¡erentiation in the rainbow trout. Furthermore, during the course of this study we also found some other type of DM-domain sequences in rainbow trout, and since trout is known to have multiple gene families due to a more recent tetraploidization event
0167-4781 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 7 8 1 ( 0 0 ) 0 0 1 8 6 - X
BBAEXP 93447 28-8-00
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
[16], this led us to analyze these new sequences along with other vertebrate DM-domain sequences. 2. Materials and methods 2.1. Animals All-female and all-male populations of rainbow trout were obtained by means of fertilizing normal eggs with sperm of phenotypic XX or YY males. The YY male genotype is viable in rainbow trout [13] and these XX and YY males were produced as previously described [17]. Steroid treatments were carried out at the INRA experimental ¢sh farm (Drennec, France), by dietary administration starting from the onset of the ¢rst feeding, i.e., reabsorption of the vitellin vesicle at 55 days post-fertilization (dpf). 11L-Hydroxyandrostenedione (11LOHv4, 10 mg/kg of food) and 17L-estradiol (E2, 20 mg/kg of food) (Sigma) were added to the food (dry pellet food, Aqualim) in ethanolic solution with 40 ml of ethanolic solution per kg of food. For tissue sampling, ¢sh were killed either by decapitation or by a prolonged overdosed anesthesia in 0.5% 2phenoxy-ethanol. For histological analysis gonads were ¢xed in Bouin-Holland £uid, dehydrated in aqueous ethanol 70% and 95%, then in ethanol/butanol (5:95) and ¢nally butanol. Tissues were embedded in para¤n, and 5Wm sections were stained with Regaud hematoxylin [18]. For the semi-quantitative reverse transcriptase^polymerase chain reaction (RT^PCR) assay in rainbow trout, 200^400 gonads were dissected out under the dissecting microscope at weekly intervals from 55 to 127 days post-fertilization (dpf). Gonads were immediately frozen in liquid nitrogen and stored at 380³C until RNA extraction. For the analysis of natural gonadal di¡erentiation, gonads were sampled in all-male (1830 gonads) and all-female (1860 gonads) populations at around 55 dpf ( þ 5 dpf). For the study of the androgen- and estrogen-induced di¡erentiation, gonads were sampled at 65 dpf ( þ 2 dpf) in all-female control populations (785 and 755 gonads), allmale control populations (752 and 781 gonads), all-female populations treated with 11LOHv4 (10 mg/kg of food) for 10 days from the onset of the ¢rst feeding (905 and 845 gonads) and all-male populations treated with E2 (20 mg/ kg of food) for 10 days from the onset of the ¢rst feeding (778 and 500 gonads). Stages of rainbow trout spermatogenesis were determined by histological analysis according to Billard and Esca¡re [19]. Stage I corresponds to immature ¢sh, stages II and III to early spermatogenesis, stages IV and V to mid-spermatogenesis and stage VI to spermiation. 2.2. cDNA cloning Degenerate primers were designed (underlined with dot-
181
ted lines in Fig. 2) according to a nucleotide alignment of di¡erent DM domains including those of D. melanogaster dsx, C. elegans mab-3, and human DMRT1 (sense primer sequences, DMRT51: TCSCCIMGSMYRCCSAAITGYGC and DMRT52: GCCTCRCCGYTIAARGGICACAA; antisense primer sequence, DMRT31: CYMARRGCIACYTGIGCIGCCAT). Touchdown PCR was carried out as described [20], using cDNAs reverse transcribed from mRNA (1 Wg) originating from various rainbow trout tissues (testis, ovary, brain, heart, muscle, liver and kidney) and genomic DNA. PCR products of the expected size were then directly subcloned into PCR2.1 vector (InVitrogen). A PCR fragment, homologous to DMRT1, was then used as a probe to screen a VZAP rainbow trout testis cDNA library according to standard methods [21], and the positive clones were ¢nally sequenced on both strands. 2.3. Sequences analysis Phylogenetic reconstruction was made with PHYLO_ Win [22] using the Neighbor-Joining Method [23] on amino acid sequences using Poisson correction and a global gap removal option. Two other methods, namely Maximum Parsimony (also implemented in PHYLO_Win) and a Quartet Maximum Likelihood [24], were also used. Bootstrap analysis with 1000 replicates was used to assess strength of nodes in the tree [25]. Accession numbers of the sequences are as follows : Gallus DMRT1, AF123456; Oncorhynchus DMRT1, AF209095; Danio DMRT2 (TERRA), AF080622; Oncorhynchus DMRT2, AF209096; Oncorhynchus DMRT4, AF209097; Tetraodon DMRT4, AJ251455; Human DMRT3, AF193872; Tetraodon DMRT3, AJ251456 ; Drosophila DSX, M25292; Bactrocera DSX, AF029676; Caenorhabditis MAB-3, U49831. The Homo sapiens chromosome 1 sequence containing a human DMRT4 homologue is referenced AL049637. 2.4. Semi-quantitative RT^PCR analysis Semi-quantitative RT^PCR was performed as previously described [17]. Brie£y, Poly(A) RNA were isolated directly from frozen gonads using Dynabeads (Dynal, Oslo, Norway) as described by the manufacturer. After reverse transcription, standard PCR reactions including negative (water only) and genomic DNA controls (with a low optimized number of cycles, i.e., 20 cycles for L-actin and 15 for rtDMRT1) were performed with speci¢c primers for each gene of interest i.e., rtL-actin (rtL-actinsense, 5P-AAAGACCCTGAGTTCATCATGC-3P; rtLactin-antisense, 5P-CCCAGTCTCCACTAATCCCA-3P), rtGAPDH (rtGAPDH-sense: 5P-GACCTCTGTGTTGGAATCAACG-3P; rtGAPDH-antisense, 5P-TCCCATTGCAGCTTTCTGGAGC-3P), and rtDMRT1 (underlined with double lines in Fig. 2, rtDMRT1-sense, 5P-TTCCTG-
BBAEXP 93447 28-8-00
182
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
Fig. 1. (A) Phylogenetic tree connecting some `DM domain' sequences and Clustal alignment of these DM-domain sequences. Numbers in each branch represent the bootstrap values obtained by Neighbor-Joining, Maximum Parsimony and Maximum Likelihood, respectively. Branch lengths are proportional to evolutionary distances. (B) Amino acid sequences are compared to human DMRT1 gene. Dots indicate identical residues and gaps are represented by dashes. The sequence of Caenorhabditis elegans C34D1.1 related to MAB-3 gene is not represented but is used as a root for the construction of the phylogenic tree. C. elegans MAB-3b corresponds to the second DM domain of the MAB-3 gene.
CAACCGGCCAGAAG-3P; rtDMRT1-antisense, 5P-AACAACCTCCTGACTGGACA-3P). One Wl of each PCR reaction was dotted on nylon membrane and hybridized with the corresponding [K-32 P]dCTP-labeled probe. Membranes were then quantitatively analyzed using an instant imager (Packard) and data were expressed as logarithms of [rtDMRT1]/[L-actin] ratios. In our standard PCR conditions (94³C for 1 min, 58³C for 1 min, 72³C for 30 s), PCR reactions always ampli¢ed a single band of the expected size and no ampli¢cation product
was detected either in our water or genomic DNA controls. 2.5. Northern and virtual northern analysis Northern blots were essentially performed as described [21], except that hybridization was performed in ULTRAhyb solution (Ambion). Due to the minute amount of tissue and hence RNA obtained from the 1800 di¡erentiating gonads, techniques such as a standard Northern blot
BBAEXP 93447 28-8-00
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
183
Fig. 2. Nucleotide sequence of rainbow trout, Oncorhynchus mykiss, rtDMRT1, with the corresponding amino acids translation represented in bold type. Start and stop codon are in bold italics and the DM domain, as previously described in [5], is underlined. The sequences underlined with dots correspond to the amino acids used for the design of the degenerate primers. The sequences underlined with double lines correspond to rtDMRT1 sense and antisense primers used for semi-quantitative RT^PCR.
and RNase protection assays were not possible because they were not sensitive enough. Thus more sensitive techniques should be applied such as virtual Northern. This technique has been proved to be absolutely feasible and sound for quanti¢cation purposes [26,27]. Virtual Northern was carried out according to the manufacturer (SMART PCR cDNA Synthesis kit, Clontech) and as previously described [17], using 1 Wg of total RNA. cDNA were loaded onto 1% TBE/agarose gels and transferred to nylon membranes (roughly from 200 to 600 ng for Fig. 3 and between 1 and 2 Wg for Fig. 4). The membrane was then hybridized in ULTRAhyb1 solution with
[K-32 P]dCTP-labeled probes. The DMRT1 probe consisted of the entire coding sequence of rtDMRT1 and reprobing of the membrane with glyceraldehyde-3-phosphate dehydrogenase cDNA (rtGAPDH, accession number AF027130) or L-actin was used as a control. 3. Results 3.1. DM-domain sequences The cloning of a DMRT1 homologous cDNA in rain-
BBAEXP 93447 28-8-00
184
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
Fig. 3. Expression of rtDMRT1 during gonadal di¡erentiation of rainbow trout. (a) Virtual Northern blot analysis of rtDMRT1 and rtGAPDH gene expression in rainbow trout di¡erentiating gonads. Increasing amounts of cDNA were loaded onto the gel (4, 6, 8, 10 and 12 Wl of the PCR reaction, at a concentration around 50 ng/Wl). (b) Histological cross-section of the gonads sampled at 60 dpf showing an absence of meiotic oocytes. Gc, gonia cells; Ic, interstitial cells; Bc, blood cells; Mes, mesogonium. (c) RT^PCR semi-quantitative analysis of rtDMRT1 (represented as mean þ S.D.) of the ratio over L-actin with an arbitrary logarithmic scale), before (pre-di¡erentiation), during (di¡erentiation), and after (post-di¡erentiation) histological differentiation. M, male ; F, female. N represents the number of samples analyzed.
bow trout was performed using RT^PCR and degenerate primers designed against conserved regions of di¡erent DM domains. Three types of sequence showing high levels of sequence identity with di¡erent DM domains were found in rainbow trout (in testis, ovary and genomic DNA). The relationships existing between these various sequences (Fig. 1) indicated that one of them (rtDMRT1 found only in testis, primers DMRT52UDMRT31) clearly clustered with the human and chicken DMRT1 cDNA sequences. A second DM homologue domain (found in ovary, primers DMRT51UDMRT31) is clearly the homologue of the zebra¢sh, Danio rerio, TERRA [28] and human DMRT2 [5,28] genes (rtDMRT2). Finally, the third type of DM domain homologous sequences (found both in ovary and genomic DNA, primers DMRT51UDMRT31) is related to a human gene found in the sequence of chromosome 1, that we tentatively named DMRT4 (accession number AL049637). We also isolated two sequences from another ¢sh species, the pu¡er¢sh, T. nigroviridis, and found that these additional sequences (pDMRT4 and pDMRT3) are related to human and rainbow trout DMRT4 and also to the recently described human DMRT3 sequences.
3.2. rtDMRT1 sequence In order to look at the implication for DMRT1 in gonadal di¡erentiation, we used the rtDMRT1 PCR fragment to screen a rainbow trout testis cDNA library. Among a few positives obtained, we sequenced a 1138 bp clone (Fig. 2) encoding a predicted 325-amino-acid protein that exhibits 90% and 87% sequence identity, respectively, with chicken and human DMRT1, but this identity is only found inside the DM domain. Outside this domain, no important sequence similarity was detected and especially the 3P end proline/serine rich region described in other species was not found in the rainbow trout sequence. 3.3. rtDMRT1 expression during gonadal di¡erentiation Semi-quantitative RT^PCR and virtual Northern analysis were carried out on genetic mono-sex populations to analyze rtDMRT1 gene expression during gonadal di¡erentiation. rtDMRT1 was shown to be highly expressed in the di¡erentiating testis when compared to the di¡erentiating ovary (Fig. 3a,c). Virtual Northern hybridization
BBAEXP 93447 28-8-00
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
with rtDMRT1 probe revealed one major band around 1.5 kb and two faint additional signals in the di¡erentiating testis (represented by arrows in Fig. 3a). Only the upper two bands were expressed in female di¡erentiating gonads but at an extremely low level. Histology of the gonads of the all-male and all-female populations sampled for the virtual Northern analysis is presented in Fig. 3b. This analysis con¢rmed that these gonads were still histologically undi¡erentiated at that time, i.e., without any sign of meiotic oocyte (as the early detection of oocyte meiosis in females is the main histological criteria used for assessing gonadal di¡erentiation in ¢sh [30,31]). By semi-quantitative RT^PCR, we also con¢rmed this early testicular expression and shown that this expression was sustained until the ¢rst male meiosis around 120 dpf (Fig. 3c).
185
Fig. 5. Expression of rtDMRT1 (top panel) in some adult tissues of rainbow trout by Northern blot. Two Wg of mRNA was loaded onto each lane from the following tissues: T, testis ; O, ovary; S, skin; H, heart; L, liver ; pK, posterior kidney; aK, anterior kidney. Reprobing with L-actin (bottom panel) was used as a loading control.
Virtual Northern analysis (Fig. 4) was also carried out on di¡erentiating gonads sampled from genetic mono-sex populations treated either with estrogen (E2, 20 mg/kg of food) or androgens (11LOHv4, 10 mg/kg of food). In the estrogen-treated all-male population, rtDMRT1 was found after 10 days of steroid treatment to be decreased 2.5-fold when compared to the all-male untreated population. Conversely, expression was not detectable, at short exposure time, in both androgen-treated and untreated control all-female populations. After an over exposition of this virtual Northern blot, a slight rtDMRT1 expression was detected but without any important di¡erence between treated and untreated female populations. 3.4. rtDMRT1 expression in some adult tissues In adult, rtDMRT1 expression was detected by Northern blot in the testis, with a single 1.5 kb transcript (Fig. 5), but was not detected in the other tissues tested (ovary, heart, intestine, skin, kidney and liver). Using RT^PCR analysis (Fig. 6), we con¢rmed an important expression in the testis but we also detected a slight expression in the previtellogenetic ovary. No expression was detected
Fig. 4. Expression of rtDMRT1 during steroid-induced gonadal di¡erentiation in rainbow trout by virtual Northern blot analysis (each lane contains about 1^2 Wg of cDNA). Reprobing with L-actin (bottom panel) was used as a loading control. (a) Expression of rtDMRT1 following a feminizing estrogen treatment (17L-estradiol, 20 mg/kg of food) on an all-male (M), ME2: estrogen-treated male population. (b) Expression of rtDMRT1 following a masculinizing androgen treatment (11L-hydroxyandrostenedione, 10 mg/kg of food) on an all-female (F), F.11Lv4: androgen-treated female population. The right panel is an over-exposition of the virtual Northern shown in the left panel.
Fig. 6. (Top panels) Expression of rtDMRT1 in some adult tissues of rainbow trout by RT^PCR. RT^PCR was carried out with 25 or 30 cycles and the 25-cycle experiment was transferred on nylon membrane and hybridized with an rtDMRT1 probe. T, testis; O, ovary; B, brain; H, heart; L, liver ; M, muscle; aK, anterior kidney; pK, posterior kidney. RT^PCR also included negative (3) and genomic DNA (G) controls. RT^PCR with rtGAPDH primers (bottom panel) was used as a control.
BBAEXP 93447 28-8-00
186
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
4.2. rtDMRT1 expression.
Fig. 7. (Top) Expression of rtDMRT1 during rainbow trout spermatogenesis by Northern blot analysis. About 20 Wg of total RNA was loaded onto each lane. Reprobing with RNA 28S was used as a loading control. The ratio of rtDMRT1 over RNA 28S (bottom) was estimated following quanti¢cation of rtDMRT1 and RNA 28S signal using an instant imager.
by RT^PCR for the following tissues: brain, heart, liver, muscle, and kidney (Fig. 6). Expression analysis by Northern blot during spermatogenesis (Fig. 7) revealed that rtDMRT1 is highly expressed at comparable levels in immature testis (stage I), early spermatogenesis (stages II and III) and mid spermatogenesis (stages IV and V). However, during spermiation (stage VI) this expression is strongly reduced (to less than 10% of the average of other stages). 4. Discussion 4.1. DM-domain gene family We have isolated a total of four di¡erent DM-domain genes in ¢sh. Their homology within their DM domain clearly identi¢es them as members of the DMRT family (DMRT stands for Doublesex and MAB-3 Related Transcription factor according to the proposed terminology [5]). The phylogenetic analysis reveals that these sequences belong to four di¡erent genes. The full-length sequence from trout described in this paper is clearly a homologue of the human and chicken DMRT1. These four DMRT genes are found in at least mammals and ¢shes (e.g., [1,5,7,28]), and thus it can be inferred that they result from an ancient duplication that arose very early in vertebrate evolution as has been found for other gene families [32]. Considering the chromosomal location of these genes in human, the most divergent gene in our analysis, DMRT3, is the only one to be located on human chromosome 1, whereas the three others are clustered in chromosome 9 [6]. This may suggest that they arose by tandem duplication, a view that is not in agreement with the hypothesis of four paralogous genes common to all vertebrates.
DMRT1 is now well known as a transcription factor likely to be involved in sex di¡erentiation in many vertebrates [7^12]. Our results clearly demonstrate, for the ¢rst time in ¢sh, that rainbow trout DMRT1 (rtDMRT1) is also highly expressed during testicular di¡erentiation but not during ovarian di¡erentiation. This testicular expression is detected as early as 2^3 weeks before the ¢rst histological criteria of sex di¡erentiation in ¢sh that is the early oocyte meiosis detection in di¡erentiating ovaries [30,31]. Expression during steroid-induced di¡erentiation demonstrate that a short-term feminizing treatment with estradiol-17L is able to downregulate the expression of rtDMRT1. This downregulation by a feminizing agent is in agreement with the idea that DMRT1 is really a testicular di¡erentiation gene and, in a similar manner, masculinizing temperature has been found to increase DMRT1 expression in a temperature-dependent sex determination (TSD) species of turtle [12]. However, in rainbow trout this downregulation by estrogens was found to be only partial and we did not ¢nd any stimulation of rtDMRT1 expression following masculinization of a genetic female population with androgens. This may likely be due to our sampling date that is only 10 days after the beginning of the treatment. Such short-term e¡ects of steroid treatments on some gene transcriptions have been already shown during gonadal di¡erentiation in rainbow trout (Govorun et al., unpublished results) and in the case of rtDMRT1 this absence of important e¡ects may indicate an indirect action of these steroids on the regulation of DMRT1 gene expression. This could suggest that DMRT1 is not a direct downstream gene under the control of steroids, and may have a more upstream position in the sex di¡erentiation cascade. Upstream genes involved in that cascade are generally thought to be less conserved genes [33], as is the case for instance for SRY in mammals, where only the HMG box shows a good conservation [34]. However, the non-vertebrate homologues of DMRT1, dsx in Drosophila and MAB-3 in nematodes, are thought to be more downstream genes in the cascade of sex di¡erentiation [35]. The exact position of DMRT1 in the sex di¡erentiation cascade of vertebrates remains to be elucidated, but its expression in normal testis development [8^11], and at the same time as Sry, i.e., just before testicular di¡erentiation [8], would favor an upstream position and a major role for that gene. The hypothesis of DMRT1 activating Sry has even been proposed [9], based on an earlier detection of DMRT1 gene expression in mouse (E9.5 for DMRT1 and E10.5 for Sry). However, in ¢sh, as in other non-mammalian vertebrates, no Sry equivalent has been found, and thus DMRT1 may even play a more crucial role in these species than in mammals. In rainbow trout adult tissues, the expression of rtDMRT1 was found in the testis as previously found in human [1] and mouse [9], with a transcript size of 1.5 kb,
BBAEXP 93447 28-8-00
O. Marchand et al. / Biochimica et Biophysica Acta 1493 (2000) 180^187
identical to the chicken transcript [7] but smaller that the human and mouse ones [1,10]. However, in rainbow trout there is also a small expression of rtDMRT1 in the ovary. This expression in the developing ovary is 80 times less than in the developing testis before histological di¡erentiation but this di¡erence is reduced to a factor of 40 during, and to a factor of 10 after histological sex di¡erentiation. This was also con¢rmed by RT^PCR analysis in adult tissues. Expression in developing ovaries has also been found in the chicken and alligator, although a higher expression was always detected in developing male embryos [8,9]. In rainbow trout, rtDMRT1 is highly expressed during spermatogenesis until late spermatogenesis, and decreases thereafter at spermiation. In mouse, expression of DMRT1 is localized both in Sertoli cells and primordial germ cells [9] but there is no indication as to whether expression is also sustained in germ cells at later stages such as spermatid or spermatozoa. The decrease in rtDMRT1 expression seen in rainbow trout at spermiation may thus either indicate a dilution of an important rtDMRT1 expression in Sertoli cells that are considerably diluted by germinal cells during the late stages of spermatogenesis, or a sudden decrease in the rtDMRT1 expression in spermatozoon that are the major cell type at spermiation. 4.3. Conclusion These results clearly demonstrate a sexually dimorphic expression of rtDMRT1 during gonadal di¡erentiation in rainbow trout, and this reinforces the hypothesis of an important implication for DMRT1 in testicular di¡erentiation in all vertebrates [1,5^12]. However, the occurrence of multiple genes with `DM domains' in ¢sh would also require further study since like in mammals [1,5,6,29], at least four `DM domain'-related genes exist in ¢sh, and these may also have implications for developmental processes. Acknowledgements This work was supported by INRA, and by a European Community project, PL 97-3796 (Y.G.), as well as by CNRS and ENS (V.L.). We thank Jean-Marc Vanacker, Franck Delaunay and Alexis Fostier for critical reading of the manuscript. References [1] C.S. Raymond, C.E. Shamu, M.M. Shen, K.J. Seifert, B. Hirsch, J. Hodgkin, D. Zarkower, Nature 391 (1998) 691^695.
187
[2] K.C. Burtis, B.S. Baker, Cell 56 (1989) 997^1010. [3] M.M. Shen, J. Hodgkin, Cell 54 (1988) 1019^1031. [4] R. Veitia, M. Nunes, R. Brauner, M. Doco-Fenzy, O. Joanny-Flinois, F. Jaubert, S. Lortat-Jacob, M. Fellous, K. McElreavey, Genomics 41 (1997) 271^274. [5] C.S. Raymond, E.D. Parker, J.R. Kettlewell, L.G. Brown, D.C. Page, K. Kusz, J. Jaruzelska, Y. Reinberg, W.L. Flejter, V.J. Bardwell, B. Hirsch, D. Zarkower, Hum. Mol. Genet. 8 (1999) 989^996. [6] C. Ottolenghi, R. Veitia, L. Quintana-Murci, D. Torchard, L. Scapoli, N. Souleyreau-Therville, J. Beckmann, M. Fellous, K. McElreavey, Genomics 64 (2000) 170^178. [7] I. Nanda, Z. Shan, M. Schartl, D.W. Burt, M. Koehler, H. Nothwang, F. Grutzner, I.R. Paton, D. Windsor, I. Dunn, W. Engel, P. Staeheli, S. Mizuno, T. Haaf, M. Schmid, Nat. Genet. 21 (1999) 258^259. [8] C.A. Smith, P.J. McClive, P.S. Western, K.J. Reed, A.H. Sinclair, Nature 402 (1999) 601^602. [9] C.S. Raymond, J.R. Kettlewell, B. Hirsch, V.J. Bardwell, D. Zarkower, Dev. Biol. 215 (1999) 208^220. [10] A. De Grandi, V. Calvari, V. Bertini, A. Bulfone, G. Peverali, G. Camerino, G. Borsani, S. Guioli, Mech. Dev. 90 (2000) 323^326. [11] B. Moniot, P. Berta, G. Scherer, P. Sudbeck, F. Poulat, Mech. Dev. 91 (2000) 323^325. [12] J.R. Kettlewell, C.S. Raymond, D. Zarkower, Genesis 26 (2000) 174^ 178. [13] B. Chevassus, A. Devaux, D. Chourrout, B. Jalabert, J. Hered. 79 (1988) 89^92. [14] T. Yamamoto, in: W.S. Hoar, D.J. Randall (Eds.), Fish Physiology, Academic Press, New York, 1969, pp. 117^175. [15] G.A. Hunter, E.M. Donaldson, in: W.S. Hoar, D.J. Randall (Eds.), Fish Physiology, vol. IX. Reproduction, Academic press, New York, 1983, pp. 223^303. [16] J. Wittbrodt, A. Meyer, M. Schartl, BioEssays 20 (1998) 511^515. [17] Y. Guiguen, J.F. Baroiller, M.J. Ricordel, K. Iseki, O.M. McMeel, S.A.M. Martin, A. Fostier, Mol. Reprod. Dev. 54 (1999) 154^162. [18] M. Gabe, Techniques Histologiques, Masson and Cie, Paris, 1968. [19] R. Billard, A.M. Escra¡re, Bull. Fr. Pisci. 256 (1975) 111^118. [20] H. Escriva, R. Sa¢, C. Hanni, M.C. Langlois, P. Saumitou-Laprade, D. Stehelin, A. Capron, R. Pierce, V. Laudet, Proc. Natl. Acad. Sci. USA 94 (1997) 6803^6808. [21] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning : A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989. [22] N. Galtier, M. Gouy, C. Gautier, Comput. Appl. Biosci. 12 (1989) 543^548. [23] N. Saitou, M. Nei, Mol. Biol. Evol. 4 (1987) 406^425. [24] K. Strimmer, A. Von Haeseler, Mol. Biol. Evol. 13 (1996) 964^969. [25] J. Felsenstein, Evolution 39 (1985) 783^791. [26] W.O. Endege, K.E. Steinmann, L.A. Boardman, S.N. Thibodeau, R. Schlegel, Biotechniques 26 (1999) 542^550. [27] O. Franz, I. Bruchhaus, T. Roeder, Nucleic Acids Res. 27 (1999) e3. [28] A. Meng, B. Moore, H. Tang, B. Yuan, S. Lin, Development 126 (1999) 1259^1268. [29] C. Ottolenghi, R. Veitia, M. Barbieri, M. Fellous, K. McElreavey, Genomics 64 (2000) 179^186. [30] J.F. Baroiller, Y. Guiguen, A. Fostier, Cell. Mol. Life Sci. 55 (1999) 910^931. [31] M. Nakamura, T. Kobayashi, X. Chang, Y. Nagahama, J. Exp. Zool. 281 (1998) 362^372. [32] V. Laudet, J. Mol. Endocrinol. 19 (1997) 207^226. [33] I. Marin, B.S. Baker, Science 281 (1998) 1990^1994. [34] P. Koopman, Cell. Mol. Life Sci. 55 (1999) 839^856. [35] W. Yi, D. Zarkower, Development 126 (1999) 873^881.
BBAEXP 93447 28-8-00