In sciarid flies (Diptera, Sciaridae), one or two paternally derived X chromosomes are discarded from the soma at early cleavages to determine the sex of the embryo (XX, females; X0, males). X chromosome(s) elimination is achieved by an abnormal anaphase segregation so that X sister chromatids do not reach the poles and are not included in the daughter nuclei. A cis-acting locus (CE) within the heterochromatin proximal to the centromere is known to regulate X chromosome elimination. By immunofluorescence analysis in early embryos from Sciara ocellaris and Sciara coprophila, we investigated histone H3 phosphorylation at Ser10, Ser28 and Thr3 prior to, and during, the X elimination process. We found that the regular syncytial nuclear divisions are characterized by a gradual loss of H3S10 phosphorylation along the chromosome arms at anaphase. Importantly, the eliminating X chromosomes show a retardation in anaphase chromatid segregation and high levels of H3S10 phosphorylation in the chromosome arms. In the present study, we provide the first evidence linking the hyper-phosphorylated H3 status of the X chromosome with a delay in sister chromatid separation at anaphase. Our findings support the idea that the CE induces a deficiency in H3 dephosphorylation in the paternal X chromosomes to be eliminated.
An outstanding example of imprinted unorthodox chromosome segregation is found in sciarid flies (Diptera, Sciaridae) where paternally derived whole chromosomes are selectively eliminated from the genome at different times during development (Metz, 1925; Metz, 1926a; Metz, 1926b; Dubois, 1932; Dubois, 1933; Metz, 1933; Berry, 1939; Rieffel and Crouse, 1966) (reviewed by Gerbi, 1986). In Sciara, during the early syncytial divisions that follow the zygote constitution (AmApXmXpXp), a first elimination event occurs when one or two paternally derived X chromosomes (Xp) are discarded from all somatic cells depending on the future sex of the embryo (XX females or XO males, respectively). Maternally supplied factors, yet undetermined, regulate the number of Xp chromosomes that are eliminated (Metz, 1938; Sánchez and Perondini, 1999). At later embryonic stages, a single Xp chromosome is discarded from all germ nuclei in both sexes. Finally, at late larval stages, the whole paternal chromosome complement is discarded during male meiosis I. As a consequence, only maternal chromosomes are retained in the sperm nucleus, and these chromosomes will be then recognized as paternal after fertilization. Another relevant feature of Sciara male meiosis, is that at meiosis II the maternal X chromosome (Xm) remains undivided, both X chromatids are included in the future sperm nucleus, and this results in the unique 3X constitution of the zygote (reviewed by Gerbi, 1986; Goday and Esteban, 2001). Importantly, studies on S. coprophila translocations and chromosome identification in S. reynoldsi led to the conclusion that it is the paternally derived non-disjoined Xm chromosome from male meiosis II that is later eliminated from the soma and germline of the embryo (Crouse, 1943).
Somatic Xp chromosome elimination in early embryonic cleavage embryos has been described previously in S. ocellaris and S. coprophila (Metz, 1925; Metz and Moses, 1926; Dubois, 1932; Dubois, 1933; Metz, 1938; de Saint-Phalle and Sullivan, 1996) (reviewed by Gerbi, 1986; Goday and Esteban, 2001). In S. coprophila, there are additional germline-limited ‘L’ chromosomes (mostly heterochromatic) that are also discarded from the embryo soma by a similar elimination process to that of Xp chromosomes (Rieffel and Crouse, 1966; Amabis et al., 1979).
The precise cellular/molecular mechanisms underlying the differential segregation behaviour of Sciara chromosomes during Xp somatic elimination have not been conclusively elucidated. The cytological picture of Xp chromosomes elimination in sciarid flies is that of “anaphase delay”. The chromosomes begin their movement towards the poles at anaphase but they remain behind the others and are not included in the daughter nuclei at telophase stage (Dubois, 1932; Dubois, 1933; Metz, 1938). The most typical example of Xp elimination is represented in Fig. 1 where, at anaphase stage, the two acrocentric Xp chromosomes are seen undivided at the equatorial plate, in contrast to the rest of the chromosomes already at the poles (Dubois, 1932). Early cytological work in S. coprophila suggested that the centromeres of the discarding Xp and L chromosomes interact with spindle fibers. In view of this, it was generally accepted that modifications in the functional activity of the centromeres must occur in the chromosomes undergoing elimination (reviewed by Gerbi, 1986; Goday and Esteban, 2001). Interestingly, in several observation of S. coprophila embryos it was evident that the eliminating chromosomes have begun to split and sometimes the daughter chromatids were seen almost separated and distinctly attached to spindle fibers by the centromeres (Dubois, 1932). These observations lead to the inference that “the process of division in those chromosomes destined to be eliminated is going somewhat more slowly than normally” (Dubois, 1932).
The process of Xp and L chromosomes loss was reanalyzed in S. coprophila embryos using confocal microscopy and tubulin antibodies to visualize the spindle (de Saint-Phalle and Sullivan, 1996). The centromeres of the Xp chromosomes were identified by rDNA FISH analysis and it was shown that during the elimination cleavages they can bind spindle microtubules, begin to separate at anaphase and remain associated with microtubules until telophase (de Saint-Phalle and Sullivan, 1996). From these observations, it emerged that Xp chromosomes sister chromatids fail to separate completely at anaphase so that the eliminating chromosomes remain at the metaphase plate. These data, therefore, do not support the classic view that Xp chromosomes elimination involves changes in centromere function. Instead, it was proposed that a failure in sister chromatid separation constitutes the mechanism of Xp and L chromosome elimination in S. coprophila (de Saint-Phalle and Sullivan, 1996) (reviewed by Goday and Esteban, 2001).
Another important piece of information refers to a cis-acting locus, the controlling element (CE), known to regulate non-disjunction of the Xm dyad at meiosis II and Xp chromosome elimination in the embryonic soma of S. coprophila (Crouse, 1960a; Rieffel and Crouse, 1966; Crouse, 1977; Crouse, 1979; Gerbi, 1986). The CE localizes within a heterochromatic block of tandem repeats of rDNA genes that are proximal to the X centromere, but the DNA sequence comprising the CE still remains to be determined (reviewed by Gerbi, 1986). However, and importantly, translocation of the CE to an autosome provokes meiotic non-disjunction of the recipient autosome and also its elimination during syncytial embryo divisions (Crouse, 1960a; Crouse, 1960b; Crouse, 1960c; Rieffel and Crouse, 1966; Crouse, 1979; Gerbi 1986; de Saint-Phalle and Sullivan, 1996). Therefore, whatever the molecular nature is of the CE, it needs to be capable of regulating two different events in the X chromosome: meiotic Xm non-disjunction by centromere inactivation and Xp embryonic somatic loss by aberrant anaphase segregation (Goday and Esteban, 2001). The mechanisms leading to meiotic Xm chromosome non-disjunction have been recently analyzed in S. ocellaris and S. coprophila males (Escribá et al., 2011). A clear link was found between the under-phosphorylated status of histone H3 in the Xm centromeric region and its functional inactivation (Escribá et al., 2011). From these results, it was inferred that the CE modifies the Xm centromeric function, causing meiotic non-disjunction, by inhibiting global histone H3 phosphorylation at the centromeric chromatin (Escribá et al., 2011). In view of these findings, an interesting question is whether the normal histone H3 phosphorylation/dephosphorylation mitotic patterns are modified in the Xp chromosomes undergoing elimination at early embryonic cleavages.
Cell-cycle-dependent phosphorylation of histone H3 (H3-P) takes place in most eukaryotes and high levels of H3-P constitute a conserved hallmark of mitotic cell division. Phosphorylation of H3 at Ser10 (H3S10-P) (Gurley et al., 1978; Wei et al., 1998; Hsu et al., 2000; Giet and Glover, 2001) and of H3S28 (H3S28-P) (Goto et al., 1999), is carried out by the mitotic kinase Aurora B that is a component of the chromosomal passenger complex (CPC) (Giet and Glover, 2001; Ruchaud et al., 2007). The action of these particular kinases is required for the proper recruitment of the condensin complex and assembly of the mitotic spindle in a phosphorylated histone H3-dependent manner (Giet and Glover, 2001). These kinases are counterbalanced by the activity of type1 phosphatases (PP1) (Hsu et al., 2000; Murnion et al., 2001) and their interplay activities is thought to be the primary means of governing histone H3 phosphorylation during mitosis and promoting the proper chromosomal condensation and segregation (Hans and Dimitrov, 2001). As a general rule, condensed metaphase chromosomes attain high levels H3-P and upon exit of mitosis a global dephosphorylation of histone H3 takes place. In Drosophila early embryos (Su et al., 1998) and Sciara spermatocytes and neuroblast cells (Escribá et al., 2011), H3S10/H3S28 dephosphorylation begins at the centromeric chromosomal regions and gradually extends along the chromosome arms as anaphase proceeds. On the other hand, phosphorylation of H3T3 (H3T3-P) is carried out by the kinase Haspin and plays a direct role in kinetochore assembly and functional activity (Dai et al., 2005). H3T3-P is crucial for the localization of the CPC at centromeres and for the function of Aurora B during mitosis (Wang et al., 2010).
In this work we have analysed the cell cycle distribution of the H3S10-P, H3S28-P and H3T3-P forms in S. ocellaris and S. coprophila, prior to and during the process of Xp chromosome elimination in embryos. We describe here the comparative location and timing of each of the H3-P modifications during the regular syncytial mitotic divisions. From this analysis, we found that the regular syncytial anaphase stage is characterised by the occurrence of a gradual H3S10 dephosphorylation from the centromeres to the tips of the chromosomes. Moreover, and importantly, we show that the eliminating Xp chromosome/s display a slower anaphase segregation rate plus a lack of H3S10 dephosphorylation in the chromosome arms, in contrast with the rest of the chromosomes. Our findings strongly indicate that, in the chromosomes to be lost, a deficiency in H3S10 dephosphorylation correlates with the delay in sister chromatid separation. The results also suggest that the CE induces the absence of H3 dephosphorylation in the Xp chromosome arms.
The regular chromosomal complement in S. ocellaris and S. coprophila embryonic pre-somatic nuclei is that of nine chromosomes: three pairs of autosomes (two acrocentric and one metacentric) and three acrocentric X chromosomes (XmXpXp). In S. coprophila, ‘L’ chromosomes can vary in number (1–4) and are metacentric. As already described (Rieffel and Crouse, 1966; de Saint-Phalle and Sullivan, 1996), in S. ocellaris, Xp elimination usually occurs at the nuclear cycle 9 in embryos of both sexes. In S. coprophila, males eliminate Xp chromosomes primarily at nuclear cycles 7, 8 and 9 while females normally at cycle 9. L chromosome elimination occurs at the earlier cycles 5 and 6, in both sexes.
Histone H3 phosphorylation in Sciara embryos during the pronuclear stage and first mitotic division
The immunolocalization of H3S10-P in S. ocellaris embryos at the pronuclear stage (Fig. 2) revealed that, as expected, H3S10-P staining was detected only when chromatin condensation takes place (Fig. 2C′) and chromosomes enter the first diploid mitotic division (Fig. 2D′). Identical results were observed in S. coprophila fertilized embryos (data not shown). Moreover, based in our observations, parental pronuclei do not fuse, but instead, undergo chromosome condensation as separate entities, before they enter the first mitotic division. This suggests that a gonomeric type of fertilization occurs in Sciara, as found in many other insects including Drosophila (reviewed by Kawamura, 2001).
Histone H3 phosphorylation during the somatic divisions prior to the occurrence of chromosome elimination in Sciara embryos
We performed immunodetection of histone H3-P during the regular mitotic syncytial divisions that characterize the early cleavages in both S. ocellaris and S. coprophila. We first tested the accessibility of the H3-P antibodies in fixed whole embryos of both species and, as expected, H3S10-P labelling was detected when the nuclei entered mitosis but not during interphase (Fig. 3).
We next examined the H3S10-P distribution from prophase to telophase mitotic stages (Fig. 4). In both species, intranuclear H3S10-P staining of prophase chromosomes was clearly detected as chromosomes condense (Fig. 4A,A′,F,F′) and the highest level of overall chromosomal H3S10 phosphorylation corresponds to metaphase stage (Fig. 4B,B′,G,G′). During anaphase, as shown in Fig. 4, H3S10-P signals were present but declined progressively starting from the centromeric regions of the chromosomes (Fig. 4C,C′) to the most distal ones (Fig. 4D,D′). At telophase, no H3S10-P signals were detectable in the newly formed daughter nuclei (Fig. 4E).
We then examined H3S28-P mitotic distribution and the results revealed that, in both species, H3S28-P was only detected when chromosomes are fully condensed and aligned in the metaphase plate (Fig. 5A,A′,C,C′). The antibody signal, moreover, is restricted to certain chromosomal sites, apparently corresponding to centromeric regions, which become significantly H3S28 phosphorylated in metaphase (arrowheads in Fig. 5A′). Therefore, in contrast with the mitotic H3S10-P pattern, H3S28-P labelling was not detected in prophase chromosomes. From these results, H3S28 phosphorylation does not appear to associate with the chromosome condensation process in the early syncytial divisions. In addition, as anaphase takes place, the segregating chromosomes are completely devoid of H3S28-P antibody staining (Fig. 5B,B′,C,C′).
The immunolocalization of H3T3-P (Fig. 6) revealed a very similar pattern to that of histone H3S28-P in both Sciara species. Prophase chromosomes were completely devoid of H3T3-P staining (Fig. 6A,A′) whereas in metaphase chromosomes, H3T3-P signals appeared in a speckled pattern apparently associated with the centromeric regions (Fig. 6B,B′). Moreover, H3T3-P antibody signals were undetectable at the initiation and progression of anaphase. Therefore, we conclude that during somatic mitosis at early cleavages H3T3 phosphorylation is only detectable at the centromeric regions of the chromosomes on the metaphase plate.
Somatic Xp chromosome elimination in Sciara embryos
Fig. 7A shows an example of S. ocellaris embryo where it is possible to visualize distinct Xp chromosomes on the end of the elimination process, when the rest of the chromosomes have already concluded anaphase segregation and have entered telophase stage. As seen in Fig. 7B, Xp chromosomes show a ‘rod-like’ appearance with the centromeres oriented towards the daughter nuclei. From our observations, a significant portion of the Xp sister chromatids have achieved anaphase separation even though they remain connected, to a variable extent, at the chromosomal ends (arrows in 7B). Based on all our data, we conclude that the elimination of Xp chromosomes involves a retardation in the anaphase progression rate, so that, sister chromatid separation takes place slower than in the rest of chromosomes.
Distribution of phosphorylated H3S10 in Sciara embryos during somatic Xp and L chromosome elimination
We next analyzed H3S10-P staining at anaphase/telophase nuclear divisions undergoing Xp and L chromosomes loss, in S. ocellaris and S. coprophila (Fig. 8). Fig. 8A–D shows typical examples of Xp and L chromosomes being eliminated and remaining at the equatorial plate. These chromosomes, unlike the rest of the chromosomes, maintain high levels of H3S10-P with the exception of the centromeric regions. At telophase stage (Fig. 8E–G), in accordance with our previous observations (Fig. 7B), Xp chromosomes have achieved a significant extent of chromatid separation. H3S10-P staining of these chromosomes (Fig. 8E′–G′) revealed that the H3S10 phosphorylation persists along the Xp chromosome arms almost to the tips of the chromosomes. From these results, we conclude that the Xp chromosomes undergoing a retardation in the segregation process exhibit high levels of H3S10-P at the chromosomal arms and, thus, do not follow the normal H3S10 dephosphorylation pattern. Therefore, when these chromosomes enter metaphase/anaphase transition, H3S10 dephosphorylation is specifically restricted to the centromeric ends of the chromosomes. This is also the case for L chromosomes that are shown in Fig. 8A′,B′, usually remain congregated at the equatorial region, most certainly due to their predominant heterochromatic nature.
The analysis of embryos at the following post-elimination cleavages showed that the discarded Xp and L chromosomes remain as cytoplasmic chromatin masses displaying H3S10-P positive staining (not shown).
In the present work, we determined the H3S10-P, H3S28-P and H3T3-P patterns during the early syncytial mitotic divisions in Sciara embryos. Consistently with previous observations in Drosophila embryos (Su et al., 1998) and Sciara meiotic cells (Escribá et al., 2011), high levels of H3S10 phosphorylation occur at metaphase chromosomes and a gradual H3S10 dephosphorylation along the chromosome arms takes place at anaphase stage. Unlike for H3S10-P, both H3S28-P and H3T3-P are restricted to centromeric regions of metaphase chromosomes and are dramatically reduced at the onset of anaphase, in contrast to what was described in other eukaryotic dividing cell types (Nowak and Corces, 2004; Xu et al., 2009; Escribá et al., 2011).
A retardation in the anaphase segregation motion leads to chromosome elimination
Classic work in Sciara embryonic development led to the view that anaphase delay underlies Xp and L chromosome elimination (Dubois, 1932; Dubois, 1933; Metz, 1938). To further characterize this process, we examined in detail Xp elimination phenotypes from the latest stages of nuclear division in whole Sciara embryos. Here, we show that by the end of telophase stage, the discarded rod-like Xp chromosomes have achieved a great extent of sister chromatid separation. In some noteworthy cases, Xp sister chromatids seem to be connected only by the telomeric chromosome ends. This particular observation could not be further analyzed because a specific Xp telomeric probe is, unfortunately, still lacking. From our observations, we conclude that even though the rate of Xp chromosome segregation is clearly reduced at the elimination cleavages, progression of Xp sister chromatid separation is feasible until the end of daughter nuclear division. Accordingly, the amount of Xp chromatid arm separation at anaphase stage is clearly less than what is seen at telophase stage. The present results, therefore, confirm early Sciara data and strongly support Dubois's inference that sister chromatid separation is much slower in the Xp chromosomes destined to be eliminated (Dubois, 1932). From all these findings, Xp chromosome elimination can be envisaged as a simple result of an individual rate reduction of sister chromatid separation. This view, on the other hand, is compatible with an abnormal Xp chromatid separation process (de Saint-Phalle and Sullivan, 1996).
In contrast to Xp chromosomes, the eliminating L chromosomes do not display sister chromatid separation and remain at the equatorial plate. This, could be due to chromatin stickiness, in view of the highly heterochromatic nature of L chromosomes DNA and the high amount of associated heterochromatic proteins (Greciano et al., 2009).
Xp and L chromosomes elimination and abnormal H3S10 dephosphorylation at anaphase
The present findings lead us to conclude that the Xp chromosomes undergoing elimination display a restriction of H3S10 dephosphorylation to the centromeric regions. Consequently, there is a failure in H3 dephosphorylation along the Xp chromosome arms at anaphase and high levels of H3S10-P are maintained from the onset of anaphase until the end of telophase. Most interestingly, the persistence of significant levels of H3S10-P is apparently unrelated with the degree of Xp sister chromatid separation that is achieved. In view of this, it is tempting to speculate that there is a direct link between the inability of Xp chromosome to dephosphorylate H3S10 along the arms and the reduction of the Xp chromosome segregation motion. Thus, it emerges that Xp chromosome elimination in Sciara associates with a failure in H3S10 dephosphorylation. This is also true for L chromosomes, which do not accomplish significant anaphase motion.
In Drosophila embryos syncytial mitosis, a localized loss of H3-P was found as anaphase stage progresses (Su et al., 1998). Cdk1 (Cyclin-dependent kinase 1) activity is required to maintain H3-P in mitosis, and proteolysis of cyclins during mitosis in Drosophila leads to a rapid H3 dephosphorylation. Based in the strict correlation between H3 phosphorylation and Cdk1 activity in Drosophila, it was suggested that a localized decline in Cdk1 activity corresponds to the localized loss of H3-P (Su et al., 1998). If this is so, the Xp chromosomes undergoing elimination in Sciara would differ from the other chromosomes in that they are able to maintain local Cdk1 activity at the chromosome arms.
A noteworthy example of chromosome elimination that exhibits some similarities with Sciara chromosome elimination process is that found in the Drosophila maternal haploid (mh) mutant (Loppin et al., 2001). At the first mh embryonic division, paternal sister chromatids fail to separate in anaphase and paternal chromatin stretches at telophase stage. Apparently, the abnormal mh phenotype segregation is not due to an irregular mitotic apparatus organization, nor to a defective centromere. Remarkably, the eliminating mh paternal chromosomes exhibit high levels of H3S10-P at anaphase and telophase stages (Loppin et al., 2001). In view of this, it was concluded that the anomalous H3-P chromatin state of paternal chromosomes is apparently incompatible with a complete sister chromatids separation (Loppin et al., 2001). Taking into consideration all the current data, it is possible that a high H3S10-P chromatin state constitutes a common requirement for the chromosome/s to be lost in early dipteran embryos.
As reviewed elsewhere (Goday and Ruiz, 2005), there are a number of insects that display programmed chromosome elimination during mitotic divisions at early embryonic cleavages. A common feature to all of them is that chromosome loss events implicate anomalous sister chromatid segregation behaviour during anaphase/telophase stages (Nicklas, 1959; Nicklas, 1960; White, 1973). Unfortunately, no H3-P studies are available in all these organisms that could, undoubtedly, shed more light in the process of chromosome elimination.
The CE and H3 dephosphorylation in the Xp chromosome
In sciarid flies, the controlling element (CE) regulates two different events in the X chromosome: meiotic Xm non-disjunction by centromere inactivation and Xp embryonic somatic loss by aberrant anaphase segregation (Goday and Esteban, 2001). As recently shown, meiotic Xm non-disjunction by centromere inactivation is achieved by the inhibition of the histone H3 phosphorylation at the centromeric chromatin (Escribá et al., 2011). It emerged that the CE governs the inability of the X centromeric chromatin to become H3 phosphorylated, unlike the rest of the chromosome (Escribá et al., 2011). The present results support the view that the Xp embryonic somatic loss by delayed anaphase segregation is achieved by the inhibition of the histone H3 dephosphorylation process at the chromosome arms, in both Sciara species. Therefore, it seems that, at the elimination cleavages, the CE governs the inability of the Xp chromosome arms to become H3 dephosphorylated at anaphase.
An understanding of the organization of the heterochromatic sequences that constitute the CE and of the mechanisms that inhibit chromatin H3 phosphorylation/dephosphorylation in cis are important questions that remain to be answered. Finally, we believe that sciarid flies would be a suitable system to analyze the different mechanisms underlying the regulation of sister chromatid segregation (cohesin, catenations, SAC, phosphatases, etc).
Materials and Methods
Fly culture and egg collection
S. ocellaris and S. coprophila were raised at 22°C as described elsewhere (Rieffel and Crouse, 1966). The duration of the different developmental stages from fertilized embryos to the elimination cleavages was determined following the timing description already established in detail (Dubois, 1932; de Saint-Phalle and Sullivan, 1996). Briefly, to analyze the early syncytial divisions, fertilized eggs were collected and fixed every 15 minutes for 3 hours. To analyze embryos prior to, and during, the process of Xp chromosomes and L chromosomes embryos were aged for 3 hours and then fixed every 1 hour for 13 hours. The number of total eggs analyzed in each experiment was not less than 300.
Sciara embryos were dechorionated and fixed in methanol following the described procedures (Goday and Ruiz, 2002). Briefly they were dechorionated with 50% Clorox for up to 2 minutes, washed in distilled water and placed in methanol/heptane (1∶1) to fenestrate the vitelline membrane by shaking the egg solution. After removing heptane, embryos were then fixed in cold methanol for at least 1 day and stored in methanol at 4°C.
Immunostaining and microscopy
Prior to antibody incubation, methanol was removed and replaced by cold acetone (−20°C) for 1 minute, the embryos were then extensively washed in PBS at room temperature. The embryos were then incubated in PBS containing 1% Triton X-100 (15 minutes) and in PBS with 3% BSA and 0.1% Tween for 1 hour at room temperature.
The primary antibodies were rabbit polyclonal anti-H3S10-P, anti-H3S28-P and anti-H3T3-P (Upstate Biotechnology) diluted 1∶60. Secondary antibodies were FITC- and Cy3-conjugated anti-rabbit (Southern Biotechnology). Secondary antibodies were diluted 1∶50 to 1∶100 for FITC-conjugated antibody and 1∶800 for CY3-conjugated antibody. Primary antibody incubation was at 4°C overnight. Secondary antibody incubation was at room temperature for at least 4 hours or at 4°C overnight. DNA was visualised with 4′,6-diamino-2-phenylindole (DAPI) staining (0.1 mg/ml; 10 minutes) and preparations mounted in anti-fading solution. Observations were made under epifluorescence optics with a Zeiss Axiophot microscope equipped with a Leica CCD camera. Digital images were processed using the Adobe Photoshop PS software.
We are grateful to A. Villasante and J. Giménez-Abián for critical comments and suggestions. We thank J. Haber for critical reading of the manuscript.
M.C.E. contributed to the experimental work and data interpretation; C.G. contributed to the conception, design, experimental work, data interpretation and manuscript writing.
This work was supported by Ministerio de Ciencia e Innovación [grant number BFU2008-02947-C02-02/BMC to C.G. and M.C.E.]