Expression profile and estrogenic regulation of anti-Müllerian hormone during gonadal development in pejerrey Odontesthes bonariensis, a teleost fish with strong temperature-dependent sex determination


  • Juan Ignacio Fernandino,

    1. Laboratorio de Ictiofisiología y Acuicultura, Instituto de Investigaciones Biotecnológicas - Instituto Tecnológico Chascomús (IIB-INTECH), Chascomús, Argentina
    Search for more papers by this author
  • Ricardo Shohei Hattori,

    1. Department of Marine Biosciences, Faculty of Marine Science, Tokyo University of Marine Science and Technology, Tokyo, Japan
    Search for more papers by this author
  • Hiroyuki Kimura,

    1. Department of Marine Biosciences, Faculty of Marine Science, Tokyo University of Marine Science and Technology, Tokyo, Japan
    Search for more papers by this author
  • Carlos Augusto Strüssmann,

    1. Department of Marine Biosciences, Faculty of Marine Science, Tokyo University of Marine Science and Technology, Tokyo, Japan
    Search for more papers by this author
  • Gustavo Manuel Somoza

    Corresponding author
    1. Laboratorio de Ictiofisiología y Acuicultura, Instituto de Investigaciones Biotecnológicas - Instituto Tecnológico Chascomús (IIB-INTECH), Chascomús, Argentina
    • Camino de Circunvalación Laguna Km. 6 (B7130IWA), Chascomús, Provincia de Buenos Aires, Argentina
    Search for more papers by this author


Pejerrey is a teleost fish presenting a strong temperature-dependent sex determination. This study was conducted to clone pejerrey amh cDNA, analyze its expression profile during thermal and endocrine manipulation of gonadal differentiation, and compare its expression with that of gonadal aromatase (cyp19a1). Amh displayed higher expression at masculinizing than at feminizing temperatures during the gonadal differentiation period. Its expression at an intermediate temperature (females 1:1 males), was high in half of the larvae and low in the other half. Cyp19a1 showed a reciprocal expression pattern to that of amh both individually- and temperature-wise. Increased cyp19a1 and amh expression was observed before morphological gonadal differentiation. Amh expression in larvae feminized by administration of estradiol or masculinized by the administration of an aromatase inhibitor was down- and up-regulated, respectively. These results show that amh plays a critical role in testicular differentiation and it is apparently modulated by estrogens in this species. Developmental Dynamics 237:3192–3199, 2008. © 2008 Wiley-Liss, Inc.


The anti-Müllerian hormone (Amh; also known as Müllerian inhibiting substance, Mis) is a glycoprotein with an extremely important function in the development of the genital tract in mammals. This hormone is responsible for inducing the regression of the Müllerian ducts in the male fetus, primordial structures that otherwise would develop into the internal reproductive organs of females (the oviduct, uterus, and the upper vagina; reviewed in Josso et al.,2006). In addition, Amh also strongly influences Leydig cell function and steroid hormone metabolism (Racine et al.,1998). Thus, Amh down-regulates aromatase gene expression, re-directing the steroidogenic pathway from estrogens to androgens (di Clemente et al.,1992; Josso et al.,1998; Rouiller-Fabre et al.,1998). Studies in birds also indicate similar functions of Amh, however they suggest a more complex interactions between Amh and steroid hormones. For instance, treatment of female bird embryos during early development with aromatase inhibitors resulted in masculinization of the gonads and this process was accompanied by an increase in amh expression (Elbrecht and Smith,1992; Vaillant et al.,2001).

Amh is also found in fish and reptiles but its role in their sexual differentiation is not nearly as well understood as in placental mammals or birds. Among fishes, only early evolved ray-finned species such as the sturgeons (Acipenser spp.) present Müllerian ducts, but these structures do not regress completely in males as they do in mammals (Wrobel,2003). On the other hand, most teleost fish do not have Müllerian ducts but they do possess an amh homologue that usually shows a sexually dimorphic expression profiles during gonadal sex differentiation and/or adulthood. For example in Japanese flounder (Yoshinaga et al.2004), zebrafish (Rodríguez-Mari et al.,2005; Wang and Orban,2007), and rainbow trout (Baron et al.,2005), amh is first expressed at low levels in the undifferentiated gonads of both sexes and then at higher levels in the male gonads compared with those of females during the process of gonadal sex differentiation. A sex-related difference in gene expression during this process was also observed in the amh type II receptor, but not in amh itself in the Japanese medaka (Kluver et al.,2007). In adult fish, amh expression patterns also vary with the species; in zebrafish, for instance, adult males and females have high expression levels (Rodríguez-Mari et al.,2005), whereas in medaka they are lower in females than males (Kluver et al.,2007). As in mammals and birds, interactions between the expression of amh and aromatase in fish are also plausible. In fact, Wang and Orban (2007) and Vizziano et al. (2007) demonstrated a reciprocal expression of amh and aromatase in the gonads of zebrafish and rainbow trout, respectively, during the gonadal sex differentiation period. Moreover, the administration of 17β-ethinylestradiol (EE2) down-regulated amh expression in zebrafish (Schulz et al.,2007). These results suggest the possibility of conserved and specific roles of Amh in different fish species, but the available information is still insufficient to allow generalizations.

Amh probably has also important functions in temperature-dependent sex determination (TSD). In two species with TSD; the reptile, Alligator mississippiensis (Western et al.,1999) and the teleost, Paralichthys olivaceus (Yoshinaga et al.,2004), increased amh expression was found at male producing temperatures. Our experimental model, the pejerrey Odontesthes bonariensis, is a gonochoristic fish with strong TSD. In this species, all female and all male populations can be obtained when the larvae are raised between hatching and the onset of histological differentiation of the gonads at low (17°C) and high (29°C) temperatures, respectively (Strüssmann et al.,1996a,1997). During the same period, intermediate temperatures (24–25°C) give rise to populations in which males and females coexist. These characteristics make of pejerrey an excellent experimental model to study the gonadal sex determination/differentiation processes in teleosts with TSD and a possible implication of Amh in these processes. In this context, the aims of the present study were to characterize the amh cDNA of pejerrey, analyze its expression profile during thermal and endocrine manipulation of gonadal sex differentiation, and compare the timing of expression with that of cyp19a1 to understand the involvement of amh in sex differentiation and TSD, and its relation with estrogen levels in this species.


Characterization of Pejerrey amh cDNA

The full-length pejerrey amh was found to have 1989 bp with a predicted open reading frame (ORF) of 1571 bp. The cDNA presented the initiation codon at position 62, the stop codon at position 1594 and the polyadenylation signal 14 bp upstream the 3′-end. This sequence is available at NCBI GenBank with the accession number AY763406. The deduced protein sequence has 510 amino acid residues and showed the highest identity with blue tilapia (60.3%), followed by Japanese flounder (59.5%) and European sea bass (58.9%; Fig. 1; Table 1). Identity was low when compared with mammalian protein sequences (20.9% against human and mouse).

Figure 1.

Alignment of Amh protein sequences. Odontesthes bonariensis (pejerrey, NCBI GenBank accession no. AY763406), Homo sapiens (human, D82968), Mus musculus (mouse, NM007445), Gallus gallus (chicken, NM205030), Paralichtys olivaceus (Japanese flounder, AB166791), Oryzias latipes (medaka, AB214971). Oreochromis aureus (blue tilapia, DQ257618), Dicentrarchus labrax (European sea bass, AM232701), Danio rerio (zebrafish, AY721604), Salmo salar (Atlantic salmon, AY722411), and Anguilla japonica (Japanese eel, eSRS21c19, AB074569). Equal residues and conserved substitutions are highlighted by black and gray background, respectively. The position of the carboxy-terminal domain (TGF-β) is shown by a dotted box.

Table 1. Amino Acid Identity (%) of the Entire Amh Protein and of the Carboxy-Terminal Region (TGF-β Domain) Between Pejerrey and the Other Species Shown in Figure 1a
Pejerrey (%) identityHumanMouseChickenJapanese flounderMedakaBlue tilapiaEuropean sea bassZebrafishAtlantic salmonJapanese eel
  • a

    The entire protein had 510 amino acid residues and low identity with that of other classes while the TGF-β domain had high identity.

Entire protein20.920.922.159.558.660.358.929.241.532.9
C-terminal domain37.637.632.371.080.474.267.752.760.249.5

Amh and cyp19a1 Expression During Thermal Manipulation of Gonadal Differentiation

The first morphological signs of gonadal sex differentiation were observed at weeks 7 and 6 after hatching for ovaries at FPT (female producing temperature; 17°C) and MixPT (mixed-sex producing temperature; 25°C), respectively, and weeks 7 and 6 for testis at MixPT and MPT (male producing temperature; 29°C), respectively. The percentages of males calculated at the end of the experiment were 0%, 55.2%, and 100% at the FPT, MixPT, and MPT, respectively.

Relative quantification of the amh transcript abundance by real time polymerase chain reaction (RQReal Time PCR) showed significant differences between larvae kept at FPT and MPT (Fig. 2). Amh expression increased slowly but steadily from the first week at both the FPT and MPT but the increase was clearly more pronounced in MPT. Expression in both thermal regimes reached peak values at week 5 and significant differences between them were observed between 5 and 7 weeks (P < 0.01). At week 8, the expression decreased and had similar levels at both temperatures. At the MixPT (25°C), amh expression was low in all individuals examined on weeks 1–4 (Fig. 3). Individuals showing increased amh expression were first detected on week 5 and by week 6 they had clearly separated in two groups with low and high values resembling either those observed at the FPT and MPT, respectively. This tentative separation of individuals with low and high expression was supported by comparison to an amh expression threshold for males of 0.0257 (defined as the mean + 2SD of the values at the FPT). Cyp19a1 showed increased expression in part of the individuals at the MixPT from week 4, a week before the first noticeable increased in amh expression. From week 5, individuals with elevated cyp19a1 expression constituted approximately half of the fish on each week and these individuals had characteristically low values of amh expression. This bimodal, reciprocal expression pattern of amh and cyp19a1 at the MixPT continued through week 7.

Figure 2.

Relative quantification of the amh transcript abundance by real time polymerase chain reaction (RQReal Time PCR) of amh expression in pejerrey larvae during gonadal sex differentiation at FPT (female producing temperature, 17°C) and MPT (male producing temperature, 29°C). Symbols and bars indicate the means and SEM, respectively. Asterisks represent statistically significant differences between rearing temperatures and arrows indicate the onset of ovarian and testicular differentiation.

Figure 3.

Relative quantification of the amh transcript abundance by real time polymerase chain reaction (RQReal Time PCR) of amh (solid squares) and cyp19a1 (open triangles) expression in individual larvae at MixPT (25°C) from week 1 to 8. Amh threshold for testicular formation (mean + 2SD of values at FPT (female producing temperature: 0.0257; dotted line).

The tissue distribution of amh expression during gonadal sex differentiation was characterized using in situ hybridization (Fig. 4). This analysis revealed positive signals only in the gonads and not in other larval tissues (data not shown), and the results broadly agreed with those from RQReal Time PCR. An important dimorphic expression was detected at the MPT compared with the FPT at week 6. The amh signals were concentrated in the medullar region of the gonads.

Figure 4.

A,C: Localization of amh expression by in situ hybridization in larval gonads at 6 weeks after hatching at MPT (male producing temperature, A) and FPT (female producing temperature, C). B,D: Hematoxylin–eosin (H&E) -stained sections for comparison. Scale bars = 10.0 μm.

Amh Expression During Endocrine Manipulation of Gonadal Differentiation

Administration of estradiol and the aromatase inhibitor Fadrozole to larvae at the MixPT during the critical time of sex differentiation produced 100% and 11.8% of females, respectively, whereas the untreated, control group, had 31.6% of females. The first evidences of morphological ovarian differentiation in the estradiol and Fadrozole treatments were observed on week 6 after hatching as in the control group. Testis differentiation was observed at weeks 7 and 10 in the control and Fadrozole groups, respectively. The estradiol-treated group showed a slight, transient increase in amh expression between 0 and 2 weeks but expression thereafter was suppressed to levels comparable to those at the FPT (Fig. 5). In contrast, most larvae treated with Fadrozole showed increased amh expression between weeks 2 and 6 and reached values comparable to those obtained at the MPT.

Figure 5.

Amh expression in groups treated with estradiol (50 μg/g of fish food; open circles) and Fadrozole (500 μg/g of fish food; gray triangles) at MixPT (25°C) between 0 and 6 weeks after hatching. The control group (no pharmacological manipulation) is indicated with black circles.


This study describes an amh ortholog in pejerrey, which shows similar size with those of other bony fishes and also the characteristic transforming growth factor-beta (TGF-β) domain. The comparison of the pejerrey Amh amino acid sequence with homologous sequences from other vertebrates evidenced low evolutionary conservation between classes. While the entire protein had low identity, the TGF-β domain, located in the carboxy-terminal region containing the characteristic seven cysteine residues responsible for the biological activity of this protein (Boyd et al.,1990), was highly conserved. Of interest, even though pejerrey does not have Müllerian ducts, amh expression showed a temporo-spatial pattern that suggests a role for this protein in the gonadal differentiation process. For instance, in this case, amh was expressed predominantly in the somatic cells of the medullar area of the gonads of putative males, which is devoid of germ cells and where eventually the efferent ducts are formed. This pattern of distribution agrees with reports for Japanese flounder (Yoshinaga et al.,2004), rainbow trout (Baron et al.,2005), zebrafish (Rodríguez-Mari et al.,2005), and Japanese eel (Miura et al.,2002) and suggests that amh plays a similar role, probably related to the formation of the efferent ducts, in these four teleost species.

The analysis of amh expression during gonadal sex differentiation of pejerrey larvae reared at different thermal regimes clearly showed a temperature dimorphic expression pattern. Thus, fish kept at the MPT, which became 100% males, showed markedly higher values than those at the FPT (100% females). An increase in amh expression (over a tentative threshold for maleness determined from the values observed at the FPT) was also observed in approximately half of the animals raised at the MixPT. Because this thermal regimen produced 55.2% males, we surmised that animals with high amh expression represent putative males (this assumption was supported by the low cyp19a1 expression in these samples—see further discussion). Our analysis also showed that a significant rise in amh expression in putative males at the MPT and MixPT occurred approximately 1 week before the appearance of the first morphological signs of testicular differentiation. Similar results have been obtained for the Alligator mississippiensis where amh expression began before testicular differentiation in embryos incubated at the MPT, while in gonads of animals kept at FPT, it could not be detected (Western et al.,1999). Despite this, we cannot conclude that Amh is the temperature-activated switch that induces testicular differentiation in pejerrey. The main reason is that amh expression levels were similar between putative males at the MixPT and MPT and between putative females at the MixPT and FPT. This indicates that temperature does not act directly on amh expression and suggests the possibility that this gene is regulated by another gene(s) or sex inducer(s) (see below). In this regard, its differential expression may be a consequence rather than the cause of gonadal sex differentiation.

The expression of several genes which appear to be fundamental for gonadal sex differentiation and development in fishes appears to be regulated by the presence of estrogens and androgens (Strüssmann and Nakamura,2002). Experimental evidences obtained in this and other studies on fishes strongly suggest that amh has complex interactions with steroid hormones. In zebrafish, for example, amh and cyp19a1 presented reciprocal expression patterns and the temporal patterns of expression suggested that the former down-regulates the latter (Rodríguez-Mari et al.,2005; Wang and Orban,2007). In another study with the same species, however, Schulz et al. (2007) demonstrated that administration of an estradiol agonist (EE2) caused down-regulation of amh and the consequent disruption of the male phenotype. As with zebrafish, both the pejerrey (this study; see also Karube et al.,2007 for cyp19a1 expression results at low and high temperatures) and rainbow trout (Vizziano et al.,2007) showed reciprocal expression profiles of cyp19a1 and amh during gonadal sex differentiation. This study further demonstrated that the onset of cyp19a1 expression at the MixPT preceded that of amh by approximately 1 week and that treatment of the larvae with estradiol during the critical time of sex differentiation consistently lowered amh expression to FPT-like levels. In contrast, treatment with the aromatase inhibitor Fadrozole caused up-regulation of amh expression and masculinization of larvae reared at the MixPT (although the ratio of males did not reach 100%), and this was likely to be associated with a reduction in estradiol levels. These evidences strongly indicate that amh expression is modulated by estradiol levels in pejerrey and indirectly implicate cyp19a1 in this regulation. Taken together with the results for zebrafish (Schulz et al.,2007) and birds (Elbrecht and Smith,1992; Vaillant et al.,2001), these results present a considerable departure from the one-way regulation of cyp19a1 by Amh demonstrated in mammals (di Clemente et al.,1992; Josso et al.,1998).

In conclusion, amh appears to be fundamental for testicular differentiation in pejerrey and, together with cyp19a1, could be used as an early molecular indicator of gonadal sex differentiation in this species. The expression patterns of amh at intermediate, sexually neutral (GSD?) and high, masculinizing temperatures (TSD) in pejerrey are similar. However, in both cases, amh is probably not the initiator of the differentiation cascade as its expression appears to be strongly modulated by estrogen levels.


Cloning of amh cDNA

For cloning of pejerrey amh, a sexually mature, adult male was anesthetized with benzocaine and killed by decapitation following internationally accepted guidelines for the care and use of laboratory animals. The testis was immediately excised and stored in RNAlater (Sigma-Aldrich, St. Louis, MO) at −80°C. Total RNA was extracted using Trizol Reagent (Invitrogen, Carlsbad, CA) and used for cDNA synthesis. Degenerate primers (amhFw and amhRv; Table 2) were designed based on highly conserved regions of amh from Anguilla japonica (NCBI GenBank accession no. AB074569), Paralichthys olivaceus (AB166791), and Salmo salar (AY722411) and used for RT-PCR. RT-PCR reactions were carried out with 1 μl of cDNA using the following program: 5 min at 95°C, 35 cycles of 95°C for 20 sec, 58°C for 20 sec and 72°C for 15 sec, and final elongation at 72°C for 3 min. To obtain the 5′ and 3′-flanking regions, cDNA was synthesized using 1 μg of RNA and the Clontech SMART rapid amplification of cDNA ends (RACE) cDNA Amplification Kit (BD Biosciences, Franklin Lakes, NJ) according to the manufacturer's instructions. Amplification of 3′- and 5′-RACE products was performed using gene-specific primers obamhFw and obamhRv (Table 2) based on the partial sequence of pejerrey amh cDNA. The amplified fragments were cloned with pGEM T Easy Vector (Promega Corporation, Madison, WI) and sequenced.

Table 2. Oligonucleotide Primers Used for cDNA Cloning, In Situ Hybridization, and RQReal Time PCR of amh, cyp19a1 (Female Molecular Marker), and β-Actin (Housekeeping Control)
PrimersSequence 5′ - 3′Amplicon
amhFwctccctgcagtcsctsccyc535 bp
obamhFwcctggcactcggcttatcctcc289 bp
ishamhFwgggacagtgatgagcagaaatgga556 bp
RQamhFwcctgtctcccgcactgttaga94 bp
RQaromFwgcgagctgtctggctgagaa100 bp
RQbactinFwctctggtcgtaccactggtatcg83 bp

Thermal and Endocrine Manipulation of Sex Differentiation

Rearing experiments on thermal and endocrine manipulation of sex differentiation were conducted separately. In both cases, fertilized eggs were obtained by artificial insemination using gametes from captive-reared broodstock and incubated at 18 ± 0.5°C in flow-through brackish water (0.2–0.5% NaCl) incubators until hatching. For the thermal manipulation experiment, approximately 400 newly-hatched larvae were stocked in three 60-L tanks set at 17 ± 0.5 °C (FPT), 25 ± 0.5°C (MixPT), and 29 ± 0.5°C (MPT). The choice of temperatures followed Str¨ussmann et al. (1997) and Ito et al. (2005). The larvae were reared in flowing brackish water, under a 16L–8D light cycle, and were fed daily to satiation with Artemia nauplii and powdered fish food (TetraMin flakes, Melle, Germany). Larvae in this experiment were sampled weekly from hatching to 8 weeks for gene expression (n = 5–10 per week per group) and for histological analyses of the time of sex differentiation (n = 10 per week per group) as described below. Additional samples were taken from each group (n = 20–40) at the end of the experimental period for histological determination of the sex ratios. Three tanks were used in the endocrine manipulation experiment and the materials and rearing conditions were similar as for the thermal manipulation except for the following details. All groups were reared at MixPT and those in two of the tanks received powdered fish food incorporated with 17β-estradiol (Wako Pure Chemical Industries, Osaka, Japan; 50 μg/g of food; Strüssmann et al.,1996b) or the aromatase inhibitor Fadrozole (Afema, Ciba-Geigy Ltd., France; 500 μg/g of food; Uchida et al.,2004). Larvae in the third tank received untreated food and were used as controls. The three groups were sampled every two weeks from hatching to week 6 for gene expression analysis and at 10 weeks for histological determination of sex ratios.

Histological Analysis of Gonadal Sex Differentiation and Sex Ratios

Fish sampled for histological analysis were anesthetized in 2% phenoxyethanol, fixed in Bouin's fixative for 24 hr, and stored in 70% ethanol. Specimens were then dehydrated, embedded in Paraplast Plus, cut in 6-μm-thick serial sections, and stained with hematoxylin–eosin. Histological preparations were examined under a microscope to identify females and males using as criteria the presence of the ovarian cavity or the efferent duct, respectively, as previously established by Ito et al. (2005). The time of histological sex differentiation for each treatment was defined as the week when ≥50% of the samples had clear ovaries or testes according to the above criteria.

Relative Quantification of Gene Expression by Real Time PCR

Larvae for gene expression analysis were anesthetized with 2-phenoxyethanol and stored in RNAlater (Sigma-Aldrich) at −80°C until further processing. Aliquots (1 μg) of RNA extracted from larval trunks as described above were treated with amplification grade Deoxyribonuclease I and reverse-transcribed using SuperScript III RNase H (Invitrogen) and oligo(dT)12–18 following the manufacturer's instructions. Primer pairs RQamhFw and RQamhRv for amh, RQaromFw and RQaromRv for cyp19a1 (accession no. EF030342), and RQbactinFw and RQbactinRv for β-actin (accession no. EF044319) were designed using the Primer Express Software (Applied Biosystems, Foster City, CA; Table 2). Reactions were conducted in 20 μl volumes containing 2× SYBR Premix Ex Taq (TaKaRa Bio Inc., Shiga, Japan), 25 ng of first-strand cDNA, and 5 pmol of each primer. The RQReal Time PCRs were performed in an ABI PRISM 7300 (Applied Biosystems). Thermal cycling conditions consisted of 1 cycle of 95°C for 10 sec and 45 cycles of 95°C for 5 sec and 60°C for 34 sec. Data were normalized against the values for β-actin and quantification by the comparative Ct method was carried out with the ABI Prism 7300 Sequence Detection Software (SDS) version 1.2 (Applied Biosystems, Foster City, CA).

Gene Expression Analysis by In Situ Hybridization

Samples for in situ hybridization were fixed in 4% paraformaldehyde and subsequently processed as for other histological preparations. In situ hybridization was carried out using a 557 bp amh probe (primers ishamhFw and ishamhRv; Table 2), labeled with the digoxigenin (DIG) RNA Labeling Kit SP6/T7 (Roche Applied Science, Penzberg, Germany). All pre- and posthybridization steps were performed using the automated Hybrimaster HS-5200 (Aloka, Tokyo, Japan). Briefly, 6-μm-thick sections were deparaffinized, dehydrated, and pretreated with proteinase K (5 mg/ml) for 7.5 min at room temperature. The reaction was stopped by washing in glycine-PBS buffer (2 mg/ml) for 10 min and the sections were then dipped in 100 mM triethylamine plus anhydrous acetic acid for 10 min. For hybridization, sections were covered with 150 μl of DIG-labeled sense or antisense RNA probe solution (1 mg/ml) and incubated overnight in a moist chamber at 60°C. After hybridization, sections were washed in 50% formamide/2× SSC and 0.1× SSC at 60°C, and then incubated with anti-digoxigenin alkaline phosphatase-conjugated antibody (1/2,000 dilution). These steps and final detection with NBT/BCIP followed the manufacturer's (Roche Applied Science) protocols.

Statistical Analysis

The data are presented as means ± standard error of the mean (SEM) in Figure 2 and as individual values in Figures 3 and 5. For comparison between the different treatments, a tentative amh expression threshold was calculated as the mean + 2SD of the expression values at the FPT between hatching and week 8 (superimposed as a dotted line in Figs. 3, 5). The significance of the differences in gene expression between temperatures in Figure 2 was determined by one-way ANOVA followed by the Bonferroni's Multiple Comparison test using the software Prism Version 4.00 (GraphPad Software Inc., San Diego, CA) Differences were considered as statistically significant at P ≤ 0.05.


We thank Tomomi Kinno for help during the rearing experiments. C.A.S. was funded by the Ministry of Education, Culture, Sports, Science, and Technology of Japan and G.M.S. was funded by the Agencia Nacional de Promoción Científica y Tecnológica (ANPCYT, Argentina).