| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | ADVANCED SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratoire de Génétique du Développement et Évolution, Institut Jacques Monod, UMR 7592 CNRS Université Paris 6 et Paris 7, 2 place Jussieu, 75 251 Paris Cedex 05, France
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionExperimental proceduresReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
In Drosophila, an unique dmlf gene is present, which was independently identified in several laboratories. First, dMLF was shown to interact both molecularly and genetically with the transcription factor DREF (DNA replication-related element factor) (Ohno et al. 2000). DREF is known to participate in the transcriptional up-regulation of many genes, especially DNA replication- and proliferation-related genes such as those encoding the DNA polymerase-
(the 180 kDa and 73 kDa subunits), dE2F1 and PCNA (Hirose et al. 1993; Takahashi et al. 1996; Sawado et al. 1998). Second, we previously identified dMLF as a physical interactor of the Suppressor of fused [SU(FU)] protein (Fouix et al. 2003), a negative regulator of the Hedgehog (HH) signaling pathway (Preat 1992; Ohlmeyer & Kalderon 1998; Ding et al. 1999; Stone et al. 1999). We showed that dmlf over-expression in the eye imaginal disc disturbed eye development and that this effect was aggravated by Su(fu) loss of function. In fly, SU(FU) inhibits the transcription factor Cubitus interruptus (CI) by sequestering it in the cytoplasm (Ding et al. 1999; Kogerman et al. 1999; Methot & Basler 2000; Wang et al. 2000). Strikingly, SU(FU) was also shown to physically interact with the chromatin remodelling protein CHD3/dMI-2 (Fouix et al. 2003), another interactor of DREF (Hirose et al. 2002; Fouix et al. 2003). Thus MLF appears to belong to a network of interactions that includes the SU(FU), DREF and CHD3/dMI-2 proteins.
In parallel, dMLF was identified in a screen for modifiers of the toxicity induced by poly glutamine (poly Q). Expansion of poly Q tracts in various proteins has been shown to be responsible for human hereditary neurodegeneration diseases such as Huntington disease and various types of cerebellar ataxia (SCA1, 2, 3, 6 and 7). Over-expression of dmlf or hMLF1 has a potent suppressive effect on poly Q induced neuronal toxicity in the retina of Drosophila transgenic flies (Kazemi-Esfarjani & Benzer 2002) and transfected rat neurons in culture (Kim et al. 2005). In the fly eye, dMLF was found to localize in aggregates containing the poly Q proteins and to reduce the recruitment of the CREB binding protein (CBP) and Hsp70 in these aggregates. It was thus proposed that dMLF could block nucleation sites and thus prevent essential cellular proteins from being recruited to the aggregates.
MLF proteins lack significant homology with any of the proteins that have already been characterized and their biochemical activity is unknown. Nevertheless, several hypotheses argue in favor of their potential role in the control of cell cycle, especially in G1/S transition. First, hMLF1 was reported to prevent erythropoietin-induced erythroid differentiation in J2E erythroleukemic cells by impeding p27Kip1-dependent cell cycle arrest (Winteringham et al. 2004). Its action leads to the stabilization of the Cullin1 (Cul1) and Skp2 components of the ubiquitin E3 ligase complex SCF (Skp2), which are involved in the proteasomal degradation of p27Kip1. However, this effect does not promote cell proliferation or decrease the G1 population. Moreover, how hMLF1 acts on Cul1 and Skp2 and how it interferes with the erythropoietin-responsive pathway, are not yet understood. Second, hMLF1 expression induces p53-dependent G1 arrest in murine embryonic fibroblasts. In this case, hMLF1 was shown to act by stabilizing p53 via the down-regulation of COP1 (Yoneda-Kato et al. 2005), a ubiquitin ligase that induces p53 degradation (Dornan et al. 2004). To have this effect, hMLF1 requires one of its physical partners, the subunit 3 of the COP9 signalosome (CSN3). Thus, depending on the system, hMLF1 can act either to prevent or to favor cell proliferation. In Drosophila, we previously reported that dmlf over-expression induces apoptosis and increased DNA replication in the wing imaginal disc (Fouix et al. 2003). Given that dMLF and DREF can associate and that DREF over-expression can induce DNA synthesis and apoptosis (Hirose et al. 2001; Yoshida et al. 2001), it is possible to speculate that the effects of dmlf over-expression are mediated by DREF. Thus, the MLF proteins could regulate cell cycle in two different ways that are difficult to connect as yet: by modulating the stability of cell cycling regulatory proteins or by regulating transcriptional activity of cell cycle genes.
In Drosophila, dMLF seems to be mainly nuclear and it is associated with chromosomes in polytenic tissues (Fouix et al. 2003). In mammalian cultured cells, MLF1 has been described as essentially cytoplasmic, but it is also present within nuclear spots (Yoneda-Kato et al. 1996, 1999; Williams et al. 1999; Lim et al. 2002). In contrast, the NPM-hMLF1 fusion is nuclear. This suggests that MLF1 subcellular localization is important for its activity and that its forced nuclear targeting resulting from its fusion with NPM could deregulate its function, leading to its leukemogenic action.
Despite these different sets of data, and due to their apparent lack of functional connection, it is as yet difficult to speculate about the function of MLF proteins and to establish a model that accounts for all of these observations. Given the conservation between fruit fly and mammalian MLF proteins, Drosophila provides a powerful model for studies that aim to elucidate the biological role of MLF proteins. To better characterize dmlf, we: (i) studied its pattern of expression, (ii) analyzed its subcellular localization both in the developing fly and in cultured cells, (iii) generated two deletions of dmlf and characterized their phenotypical effect, (iv) studied the effect of the over-expression of two dMLF variants on poly Q toxicity in a SCA3 humanized fly model and (v) analyzed its conservation by sequence analysis and by looking at the effects of hMLF1 expression in fly.
We were thus able to show that dmlf is expressed ubiquitously during embryonic and larval development and in adult gonads and that four isoforms of the dMLF protein are present that are probably the result of alternative mRNA splicing. Moreover, the level of expression of dmlf and the subcellular localization of its product can vary according to the tissue, the time of development and the cell cycle phase. Using transfected cultured fly cells, we showed that two forms of dMLF (dMLFA and B, generated by mRNA alternate splicing) differ in their subcellular localizations and that dMLF has the ability to promote the nuclear localization of the normally cytoplasmic SU(FU) protein. Next, we obtained the first dmlf mutants and showed that dmlf mutant embryos born from dmlf- mothers display a strong embryonic lethality without any obvious developmental arrest. The adult dmlf- escapers display only slight defects which are suppressed by hMLF1 expression. Finally, using an inducible neurodegeneration model, we demonstrated that dMLFA and B proteins (as well as hMLF1) can provide protection against poly Q induced toxicity at postdevelopmental stages. In conclusion, our data support the hypothesis of a functional conservation of the mlf genes from fly to human.
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
MLF-related proteins have been identified in a wide variety of organisms such as the worm Caenorabditis elegans, the chordate sea squirt Ciona intestinalis and mammals (Mus musculus and Homo sapiens). Only one mlf gene is present in nonvertebrate organisms, whereas two MLF genes, MLF1 and 2, are present in human and mouse. In fly, four alternative spliced dmlf EST have been isolated (dmlfA–D, Fig. 1A,B), corresponding to mRNA of about 1.6–1.7 kb (GENBANK accession: CG8295). All variants include the exons 3 and 4 but other short exons can be present (1, 2, 5, 6 or 7). We compared the sequence of the dMLFA spliced variant protein (309 residues) with that of different vertebrate and nonvertebrate species. Alignment of the MLF proteins (Fig. 1C) by the clustalW method revealed that MLF proteins from nonvertebrate organisms present equivalent conservation to the mammalian MLF1 or MLF2 proteins, suggesting that mammalian MLF1 and MLF2 derive from the duplication of an unique common ancestor. A high degree of sequence similarity is observed in the central region which is encoded by exon 4 and is therefore present in all dMLF variants. In this region, which contains the DREF binding region (aa 96-202), dMLF displays 49% identity and 84% similarity to hMLF1 and 48% identity and 84% similarity to hMLF2. Twenty-eight over 97 residues are strictly conserved between Drosophila melanogaster and mammals (ClustalW1.81, see the asterisks in Fig. 1C) which thus seem to be a hallmark of MLF proteins.
|
We examined dmlf expression throughout embryogenesis by in situ hybridization of whole-mount fly embryos (Fig. 2). We first attempted to perform hybridizations with different exonic regions specific to the dmlf mRNA variants A and B, B or C, but no signal could be detected in this way (data not shown), probably due to the small size of the probes, which were limited by the size of the variant specific exons (less than 200 pb). We therefore had to use an RNA antisense probe that could recognize all dmlf mRNA spliced variants (see below) and will here be collectively referred to as "dmlf RNA." dmlf RNA uniformly accumulated in the early embryo (Fig. 2A and Fig. S1). Later, after the blastoderm stage, dmlf was ubiquitously expressed, with a stronger staining in the central nervous system, the gonads, the digestive tract and two clusters of cells located in the head mesoderm and which correspond to a subset of immune cells, the crystal cells (Fig. 2B–G and Fig. S1B–C). In control mutant embryos deleted for most of the dmlf sequence (see below) no signal could be detected, indicating the specificity of the probe (Fig. 2H,I).
|
In parallel, we also examined the subcellular localization of dMLF proteins. In early blastoderm embryos prior to cellularization, dMLF proteins were largely cytoplasmic, whereas, during the next four to five cell cycles, they progressively accumulated in the nuclei at the cytoplasm's expense (Fig. 3A–C).
|
|
dMLF was also detectable in the gonads of third instar larvae (Fig. 4E,H). As shown by colabelling with the adducin-like protein Hu li tai shao (Hts) which strongly accumulates in the cytoplasm of germ line cells (Yue & Spradling 1992), dMLF specifically accumulated in the nuclei of both female germ cells (located in the central region of the larval ovary) (Fig. 4E) and male germ cells (spread throughout the larval testes) (Fig. 4H). In the adult ovaries (Fig. 4F,G), dMLF was present in the nucleus of both germ and somatic cells, from the early germarial region onwards. In adult testes (Fig. 4I), dMLF was present in nuclei of early germ cells and spermatocytes, but not in cells at later stages of meiosis or differentiation.
In summary, as its partners DREF and Su(fu), dmlf is dynamically expressed throughout development in a wide variety of tissues. This could reflect the possible general cellular function of dMLF and/or its involvement in many developmental processes. In most cells, the dMLF proteins are nuclear but can also be found in the cytoplasm at specific developmental stages, suggesting that dMLF's subcellular localization is regulated and is perhaps important for its function.
Expression and differential subcellular localization of dMLF isoforms
Immunoblotting of protein extracts from 0 to 24 h old embryos with our anti-dMLFA polyclonal antibody revealed the presence of four major bands of 33, 37, 40 and 47 kDa (Fig. 1D and Fig. S2). All these bands correspond to dMLF proteins since they are absent when the preimmune serum was used (data not shown) and, most significantly, in embryos deleted for the dmlf gene (see Fig. 1D). According to their size, these four dMLF proteins could be encoded by the four mRNA spliced variants. Moreover, dmlfA and dmlfB over-expression led to a strong increase in the intensity of the 37 and 33 kDa products, respectively (see Fig. 7D). Nevertheless, our data differ from those presented in a previous report by Ohno et al. (2000) who identified only a single embryonic protein of 38 kDa. Such a discrepancy is probably due to the use of two different antibodies that might recognize different epitopes.
|
All the published data concern dMLFA or B variants, and to compare their subcellular localizations we tagged them with fluorescent proteins (GFP or RFP) and produced them in Drosophila cultured cells. Since dMLF may be involved in the HH pathway, we used Cl-8 cells, which are known to respond properly to HH (Chen et al. 1999). As shown in Fig. 5, dMLFA-RFP is mainly nuclear with a higher accumulation in nuclear speckles (Fig. 5A), whereas dMLFB-RFP localizes both in the nucleus and diffusively in the cytoplasm (Fig. 5B). The A and B isoforms differ in their C-terminal region, with four additional residues after the conserved central region in the B form and 40 for the A form (Fig. 1B). Thus the C-terminus of dMLFA, but not of dMLFB, contains a region rich in lysine and arginine residues. This feature is characteristic of nuclear localization sequences (NLS) and one possibility is that this region targets dMLFA to the nucleus. Co-expression of the two dMLF (dMLFA-GFP and dMLFB-RFP) variants did not seem to affect their respective localizations (Fig. 5D).
|
In conclusion, our data show that four different isoforms of dMLF are expressed during embryonic development and that dMLF isoforms A and B differ slightly in their subcellular localizations, but can both induce a partial nuclear localization of SU(FU).
dmlf loss of function leads to high embryonic lethality but only to weak adult defects
So far, all analyses of the activity of dmlf have been based on its over-expression in transgenic flies, since no dmlf mutant was available. To generate a dmlf deletion by excision of P element, we used a strain (EU2490) that had a P element inserted in the first intron of dmlf (Kazemi-Esfarjani & Benzer 2002). EU2490 displayed a wild-type phenotype and the phenotype of a dmlf loss of function was unpredictable. Therefore, we screened for dmlf deletions by PCR. We thus recovered two deletion mutants (over 85 excision events) called dmlf
5-3 and dmlfC1. In dmlf5-3, the start codons and the most highly conserved central part of the different isoforms were eliminated (Fig. 1A). In dmlfC1 mutant strain, most of dmlf coding sequence was deleted (Fig. 1A). Thus, the two deletions are probably dmlf null mutants.
In the dmlf–/dmlf-– embryos born to dmlf–/dmlf-– parents, 83.3% (for dmlf5-3) and 64% (for dmlfC1) of embryos died (compared to 10% in the control mlf+/mlf+ embryos born to mlf+/mLf+ mothers) (Table 1). High levels of lethality were also observed with a dmlfC1/dmlf5-3 trans-heterozygous combination. We compared the lethality in heterozygous embryos born to mutant mothers with that of embryos born to mutant fathers. The lethality was mostly due to the lack of maternal dmlf product but it can be partially rescued by a paternal or zygotic contribution. Indeed, for heterozygous dmlf–/dmlf+ embryos lacking maternal dmlf product, the embryonic lethality is significantly reduced to 64.4% (dmlfC1/dmlf+) and 46% (dmlf5-3/dmlf+). Conversely, the levels of viability for heterozygous dmlfC1–/dmlf+ embryos born to wild-type mothers are not significantly different from those seen in wild-type embryos born to wild-type parents, whereas dmlf5-3/dmlf+ embryos have a reduced viability. The differences in the lethality of the two alleles are not understood, but it could be due to the deletion in dmlf5-3 of a sequence that is important for the expression of a neighboring gene upstream.
|
The duration of embryonic development was increased (data not shown) and the homozygous mlf mutant adults which emerged displayed subtle abnormal phenotypes (Fig. 6B,B'): ocellar bristles were absent, head macrochaetes were frequently bent or shortened and ectopic venation appeared near the longitudinal vein 2. These adult phenotypes were generally stronger in female than in males (data not shown). They were observed for both mutant strains as well as in flies carrying the dmlf deletions over chromosomes with large deficiencies covering the dmlf gene (data not shown). They were not, however, seen in the control wild-type dmlfC5 or dmlfR2 strains which were generated by a perfect P excision and therefore probably had a wild-type dmlf gene (Fig. 6A,A' and data not shown). The body size of dmlf– adults was indistinguishable from wild-type animals. Lastly, since dmlf was expressed in gonads, we checked whether the escapers were fertile and looked at the egg chambers, but no obvious defect was detectable.
|
C1 or dmlf5-3 strains are likely to be truly due to the mutation of dmlf. Given that dMLF and SU(FU) can physically interact, the wing phenotype seen in dmlf mutants could correspond to a slight deregulation of the HH pathway. No genetic interaction between these two genes could be detected, however: dmlf–/dmlf–; Su(fu)–/Su(fu)– flies display the same phenotype as dmlf–/dmlf–; Su(fu)+/Su(fu)+ mutant flies.
To summarize, we generated two deletions that encompass the dmlf gene. Their analysis provided evidence that the lack of maternal dMLF leads to a high level of embryonic lethality at various developmental stages, whereas we observed very few defects in surviving adults.
dMLF provides protection against SCA3-Q78f-induced neurodegeneration in adult flies
To analyze the effect of dMLF dosage on neurodegeneration we used the SCA3-polyQ model developed by Warrick et al. (1998). In this system, the eye targeted expression of a truncated mutant protein with a 78 CAG repeat (MJDtr-Q78) leads to a severe neurodegeneration phenotype associated with an aberrant eye morphology, reduced eye pigmentation and eye collapse due to the absence of retinal structures (Warrick et al. 1998 and see Fig. 7A,A').
To investigate whether reducing dmlf dosage could modify the effect of MJDtr-Q78, we expressed MJDtr-Q78 in dmlf –/+ flies: the eye phenotype induced by MJDtr-Q78 was neither aggravated nor rescued in this context. Next, since the dMLF isoforms A and B presented differential subcellular localization, we compared their ability to suppress the effects of MJD-Q78. To do this, we co-expressed MJDtr-Q78 with dMLFA and B isoforms (using transgenic lines UAS-dmlfA (Fouix et al. 2003) and UAS-dmlfB (Kazemi-Esfarjani & Benzer 2002)). In both cases, UAS-MJDtr-Q78/gmrGAL4; UAS-dmlfA or B/+flies showed a strong improvement in external eye structure and pigmentation (Fig. 7B,C). In comparison, eyes of gmrGAL4/+; UAS-dmlfA or B/+display no obvious phenotype (data not shown). We further examined the external morphology of the compound eyes of the UAS-MJDtr-Q78/gmrGAL4; UAS-dmlf/+flies by scanning electron microscopy (SEM) and we noticed subtle differences in the rescue provided by the two dMLF isoforms. When dMLFB was co-expressed with MJDtr-Q78, the external eye morphology was perfectly rescued but the eye was collapsed, which indicates a severe disruption of its internal morphology (Fig. 7B'). In contrast, dMLFA co-expression led to a better rescue of internal morphology since there was no collapse, despite a less regular external morphology of the eye (Fig. 7C'). Interestingly, the suppressive effect of the MLF proteins was conserved in evolution since hMLF1 expression also had the capacity to rescue the eye morphology defects induced by SCA3 (Fig. S4). In addition, we investigated the potential involvement of the dMLF partner, DREF in the suppression of MJD-Q78-induced neurodegeneration. We used transgenic lines to co-express MJD-Q78 with DREF (UAS-MJDtr-Q78/gmrGAL4; UAS-DREF/+) or a DREF dsRNA construct (UAS-MJDtr-Q78/gmrGAL4; UAS-DREFdsRNA) that reduces DREF expression (Yoshida et al. 2004). In none of these flies did we observe any improvement in the eye degeneration induced by MJDtr-Q78 (data not shown). In the case of DREF over-expression the lack of effect might simply be masked by the negative effect of DREF over-expression itself, since it has been shown to induce a rough eye phenotype (Hirose et al. 2001).
These data demonstrates a protective effect of dMLF over-expression in a SCA3 Drosophila model. However, the phenotypes observed using a gmr-GAL4 driver reflect mainly events arising during eye development preceding adult emergence. It is thus not clear whether dMLF is able to protect flies in a true neurodegenerative process occurring in adult flies. To address this question and to circumvent any developmental effect, we used a RU486 inducible "gene switch" system (Osterwalder et al. 2001) to induce a pan neural expression of MJDtr-Q78 in adult stages. This system relies on an elav-GAL4 GeneSwitch (elavGS) which expresses a conditional RU486 dependent Gal4 protein under the control of the neural elav promoter. Lifespan determination allows a quantitative assessment of statistical significance: to ensure stringent conditions, we considered the following as significant: modifications of mean lifespan greater than one third and LogRank P-value for survival curves analysis < 10–15. As shown in Fig. 8A, we observed a strong reduction in life-span of flies expressing MJDtr-Q78 in adult neurons (RU+) compared to control flies (RU–). Indeed, the former flies have a mean life-span of 8.3 days; whereas at this time 98% of control flies still survive. Expression of dMLFA or B driven in the same conditions led to a slight but significant decrease of life-span (T50 = 40 days and 27 days, respectively, Fig. S5). In contrast, decreasing the dosage of either dMLF (in dmlf–/+flies) or DREF (by co-expressing UAS-DREFdsRNA) did not modify significantly the mean lifespan of these flies (Fig. 8B).
|
Altogether these data demonstrate that over-expression of dMLF can prevent toxicity of a polyQ-SCA3 pathological protein during both developmental and adult stages. Moreover, isoform A appears to be the most effective to confer a protection against expansion of polyQ toxicity.
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
Analysis of dmlf mutant
dmlf embryos lacking maternal product show a high level of lethality, which is enhanced in zygotic mutant embryos. Nevertheless, this lethality could not be associated with a specific developmental arrest. Such embryonic lethality, associated with its ubiquitous expression, indicates that the functions of dmlf are important during the entire development of the embryo. The fact that adults that survive display only subtle phenotypes suggests, however, that dmlf is not essential after this critical period. Since dMLF binds numerous sites on the chromatin, it might participate in the transcriptional control of numerous genes, like its interactor, the transcription regulatory factor DREF, and thus be involved in various cellular and developmental processes. Similarly, mutants in genes encoding proteins involved in chromatin modification complexes (e.g. dMi-2) have also been reported to lead to a strong lethality associated with a lack of obvious specific phenotypes (Kehle et al. 1998).
Different reports indicate that dMLF might be involved in control of the cell cycle, but we did not detect any defects in either M or S phase in dmlf mutant embryos. dmlf is unique in the fly genome, but one cannot exclude that another protein might play a similar role or that an dmlf independent redundant pathway might partially limit the effects of dmlf loss of function during embryonic development. Alternatively, some transient cell cycle defects might be masked by compensatory processes, as described for the effects of DmcycE over-expression (Li et al. 1999). Indeed, the abnormally elevated cellular density induced by DmcycE is compensated by apoptosis and embryonic development lasts longer. Moreover, in this experiment, the adults that emerge are viable and fertile with no obvious abnormal phenotypes. In the case of dmlf mutants, lethality could occur at multiple stages due to a poor temporal coordination between cell proliferation and certain morphogenetic processes, and such a compensatory mechanism could account for the delayed embryonic development. The embryos that succeed in passing all the critical stages would end up forming an almost normal organism.
dmlf adults present defects in their wing venation and in their bristle morphology, which are rescued by the expression of hMLF1. Such suppression is indicative of both the specificity of these phenotypes and the conservation of the function of MLF proteins from fly to human. Bent bristles are known to occur in mutants of genes involved in actin organization. Thus, it is possible that dMLF acts on the formation or stability of actin filaments either directly (via its cytoplasmic form) or indirectly by regulating the expression of the genes that participate in the actin architecture. Such a hypothesis is reinforced by the identification, in a two-hybrid screen, of an interaction between hMLF1 and an actin-binding protein of the ankyrin superfamily (unpublished observation). Further experiments will be required to test this possibility.
Subcellular localization
The subcellular localization of the NPM-hMLF1 fusion is thought to play an important role in its pathogenicity. Indeed, this fusion protein is nuclear, whereas hMLF1 is mainly cytoplasmic (Yoneda-Kato et al. 1996). Moreover, in cultured cells, NPM-hMLF1 has a pro-apoptotic effect which is dependent upon the presence of its nuclear localization sequence (Yoneda-Kato et al. 1999). Here, we show that dMLF is mostly nuclear in fly cultured cells and in many different tissues throughout fly development, in agreement with a putative transcriptional function. We also bring several lines of evidence that indicate, however, that dMLF can also be cytoplasmic and that its subcellular localization might be regulated. Thus, in Cl-8 cultured cells, a spliced variant of dMLF (dMLF-B) is also present in the cytoplasm. Furthermore, in the eye imaginal disc dMLF is cytoplasmic in the proliferating cells, but mainly nuclear in differentiating cells, suggesting that progression through G1/S and S phases might be associated with a lower level of dMLF in the nucleus or/and that their differentiation requires its nuclear accumulation. Last, during blastoderm development, dMLF is progressively relocalized from the cytoplasm to the nucleus. A similar phenomenon was also reported with DREF (Yamaguchi et al. 1995; Hirose et al. 1996). During this transition period, two main events take place: the zygotic transcription program starts and the control of cell cycle commences profound changes associated with the introduction of the GAP phases and of the cell cycle check points. The accumulation of dMLF in the nucleus during this period suggests that dMLF could participate in these processes. Nevertheless, the differences in dMLF subcellular localization during this period of development can be explained by several non-exlusive means. Firstly, it could be due to the expression of different spliced variants, with the maternal expression of a cytoplasmic form and the zygotic expression of a nuclear form. However, our developmental Western analysis showed that the maternal dMLF corresponds mostly to the A variant, which is nuclear when expressed in Cl8 cells as well as in various types of larval tissue (data not shown). Secondly, this change in subcellular localization might result from the introduction of the G1 phase in the cell cycle. Indeed, dMLF might be excluded from mitotic chromosomes in early embryos (as it is in mitotic eye imaginal disc cells) and its re-entry in the nucleus might require the G1 phase. Lastly, the nucleo-cytoplasmic traffic of dMLF could also be regulated post-traductionally. For instance, the subcellular localization of dMLF may be controlled by its binding to 14-3-3 proteins which are known to bind and influence the activity of numerous and diverse proteins (most of which are involved in apoptosis and cell cycle regulation) in part by acting on their subcellular localization. Indeed, 14-3-3zeta has been shown to associated with phosphorylated hMLF1 (Ohno et al. 2000; Lim et al. 2002) whereas dMLF (A and B) contains a consensus sequence for the 14-3-3 binding. Nothing was reported, however, neither on the role of the interaction between 14 and 3-3zeta and hMLF1 nor on a potential interaction between dMLF and 14-3-3 in fly.
Relationship of dMLF with its partners DREF and SU(FU)
dMLF physically interacts with SU(FU), a negative regulator of the HH pathway. SU(FU) is an ubiquitously expressed protein that sequesters the transcription factor CI/Gli in the cytoplasm. Its loss of function alone has no phenotypical effect but its role can be revealed by genetic interactions with genes encoding other components of the HH pathway. dmlf mutants display no alterations that are characteristic of the HH pathway and the adult phenotypes could not be modified by Su(fu) loss of function. dMLF has the ability, however, to partially delocalize SU(FU) to the nucleus, suggesting that these two proteins can indeed associate in vivo. This, together with the fact that the majority of SU(FU) interactors previously identified are (or are likely to be) nuclear proteins (or possess a NLS) (Fouix et al. 2003), indicates that SU(FU) might have a nuclear function. Such a function has been previously reported for its mammalian counterpart, which is able to enhance the repressor effect of GLI proteins by recruiting the corepressor complex SAP18/mSin3/HDAC (Cheng & Bishop 2002). Here, the failure to identify any genetic interaction between Su(fu) and dmlf might have various causes such as functional redundancy with other genes, the fact that the effect might be restricted to specific cells (such as immunity cells since dMLF is expressed in crystal cells) or to specific conditions (such as stress). Nevertheless, our data in cultured cells also opens the possibility that dMLF's capacity to relocalize one of its partners to the nucleus, may be involved in the oncogenic effect of the nuclear NPM-hMLF1 chimeric protein.
In the nucleus, MLF proteins could play a role in transcriptional regulation. Thus, dMLF interacts with the transcription factor DREF and localizes on the polytenic chromosomes of the larval salivary glands. On the other hand, MLF proteins could also be involved in the regulation of protein degradation. Indeed, in mouse NIH3T3 fibroblasts, hMLF1 interacts with CSN3 (COP9 signalosome subunit 3) resulting in a decrease in the level of COP1 protein. This leads to p53 stabilization by impeding its ubiquitylation by the COP1 ubiquitin E3 ligase (Yoneda-Kato et al. 2005). Interestingly, all COP1 degradation substrates co-localize with it in nuclear speckles (Schwechheimer 2004). Since many targets of COP1 are transcription factors, it has been proposed that these speckles represent active sites of transcription and/or protein degradation. Because mMLF1 (Williams et al. 1999) and dMLFA are present in nuclear speckles too, MLF proteins might thus regulate transcription by controlling the stability of transcription factors such as DREF or CI.
dMLF and polyQ disease
We have shown in this paper that the over-expression of dMLF can efficiently suppress in vivo the toxicity of a pathological Q78 truncated SCA3 protein both at developmental and adult stages. These findings extend previous results obtained with polyQ models (Kazemi-Esfarjani & Benzer 2002; Kim et al. 2005). We demonstrated that the dMLFA isoform is more effective than the B isoform in suppression of polyQ-induced toxicity. Similarly, over-expression of dMLFB in the wing imaginal disc has less effect than that of dMLFA (data not shown). This dissimilarity is not understood as yet. It does not seem to be due to differences in the expression of either variant, since dMLFB accumulates at a slightly higher level than dMLFA. It might, however, be due to differences in their subcellular localization. Thus, the presence of dMLF in the nucleus and its relationship with transcription regulatory factors may be relevant to the suppression of polyQ-induced toxicity. In favor of this last hypothesis, the cytoplasmic and nuclear B isoform is less effective in suppression than the nuclear A isoform and dMLF proteins have been shown to be recruited into nuclear polyQ-positive aggregates (Kazemi-Esfarjani & Benzer 2002). The dMLF proteins present in the aggregates may limit their formation or may compete with other factors for aggregate binding sites, thus preventing their depletion within the cell, as shown for CBP in mammalian sympathetic neurons (Kim et al. 2005). Alternatively, they may play an indirect role through their direct interactions with the DREF and SU(FU) proteins. However, we could not detect any modulation of the effects of MJD-tr-Q78 expression by decreasing or by increasing the SU(FU) or DREF gene dose (this work and data not shown). Finally, an important finding is the similarity in the pathological MJDtr-Q78-induced phenotypes observed in dmlf–/+ heterozygous flies and dmlf+/+ flies. These data strongly suggest that a potential depletion of the dMLF protein through its incorporation into aggregates does not cause SCA3 pathology in flies. It would be important to check whether this conclusion is also true in different SCA mammalian models.
Conservation of MLF proteins
The MLF proteins are found throughout the animal kingdom. Furthermore, ectopic expression of hMLF1 can rescue the zygotic effect of dmlf loss of function, has a similar but weaker effect than dMLF on the development of the wing imaginal disc (data not shown) and on the reduction of polyQ-induced toxicity in the eye. Reciprocally, dMLF and its human orthologs have been shown to suppress polyQ-toxicity in primary rat neuronal cultures (Kim et al. 2005). Finally, it is worth noting that while in human, hMLF1 has been shown to be involved in hematopoietic disorders, in Drosophila, dMLF is expressed in a subset of immunity cells. These data all argue in favor of a functional conservation of this novel family of proteins and validate the use of Drosophila to decipher their normal and pathological functions.
|
Experimental procedures |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
The following sequences were aligned using the clustalW method (http://www.infobiogen.fr/services/analyseq/cgi-bin/clustalw_in.pl): Caenorabditis elegans (CeMLF), Apis mellifera (AmMLF), Drosophila melanogaster (dMLFA, from CG8295-RA, Ohno et al. 2000), Ciona intestinalis (CiMLF), Mus musculus (MmMLF1 and 2) and Homo sapiens (HsMLF1 and 2) (http://www.ensembl.org).
Drosophila stocks
EP element excision was performed on w1118; P{EP}EU2490/P{EP}EU2490 (EU2490) flies which have an EP insertion in the first intron of the dmlf gene (Kazemi-Esfarjani & Benzer 2002). dmlf5-3 and dmlfC1 were obtained in this work (see below). The dmlf+ control strains were w1118 flies or dmlfR2 and dmlfC5 (that correspond to precise excisions of EP2490) as indicated in the figures.
P{Act5C-GAL4}/TM6B (chr 3, Bloomington Drosophila Stock Center), gmr-GAL4 (chr 2, provided by Fauvarque M.O.), UAS-lacZ (chr 2, Bloomington Drosophila Stock Center), UAS-dmlfA (chr 3, Fouix et al. 2003) UAS-DREF (chr 3, Hirose et al. 1999), UAS-DREF-dsRNA (chr X, Yoshida et al. 2004), UAS-dmlfB (chr 3, Kazemi-Esfarjani & Benzer 2002), UAS-hMLF1 (M7.2 and M22.1 correspond to two independent insertions on chr 3, this work), UAS-MJDtr-Q78/gmrGAL4 (chr 2, Warrick et al. 1998), elav-GAL4 GeneSwitch (elavGS, chr 3, Osterwalder et al. 2001). Flies were kept on standard media at 25 °C, except when indicated otherwise in the legends.
dmlf mutant and transgenic lines
The EU2490 strain was first back-crossed 5 times with w1118 flies. The transposase was obtained by crossing w1118; P{EP}EU2490/P{EP}EU2490 female flies with w; Sp/CyO; P{2-3}Dr/TM6 Ubx male flies. F1 w; P{EP}EU2490/CyO; P{2-3}Dr/+ male flies were selected and crossed with w, vri7/CyO (Pw+ inserted in the vrille gene) female flies. Lines were established by crossing individual w1118; P{EP}EU2490*/CyO male flies with vri7/CyO female flies to balance excision events. Deletions were then detected by PCR. All excision points were checked by sequencing. Over 85 lines, two deletions (dmlf5-3 and dmlfC1) in the coding sequence were thus obtained, as well as two wild-type alleles (dmlfR2 and dmlfC5) which correspond to perfect excision.
P-element mediated germ-line transformation was carried out in the w1118 strain.
Evaluation of life span
Three-day-old adult male flies were collected and clusters of 30 were put in RU486-containing food tubes or control tubes without RU486. Every two days, dead flies were counted and the living flies were transferred to fresh tubes (containing the same media as the initial one). Life span was measured with at least 90 male flies per genotype and per condition except where indicated. T50 was estimated as the time at which half of flies were dead. Log rank analysis was used to assess the statistical significance of the observed variations. RU486 (Sigma) was used at 200 µg/mL.
Electron microscopy
Flies were dried overnight at 50 °C and their heads were coated with 30 nm of gold. They were then observed with a JEOL JSM 6100 scanning electron microscope.
Plasmids
The full-length hMLF1 cDNAs (Yoneda-Kato et al. 1996) were inserted between EcoRI and XhoI sites in the polylinker of the pUAST plasmid (Brand & Perrimon 1993).
For cells transfections, we built vectors encoding the complete dMLFA, dMLFB or SU(FU) proteins fused to either the Green or Red Fluorescent Protein (GFP or RFP) were built using the Gateway method (Invitrogen). The entire coding sequence (including the ATG and the STOP codons) of either dmlfA, dmlfB or Su(fu) were amplified by PCR and cloned in the entry vector pENTR/D-TOPO by directional TOPO cloning. The resulting plasmids were checked by sequencing. The destination vectors pAWG (pAct5C-GW-EGFP), pARW (pAct5C-GW-RFP) were built and given by T. Murphy.
Cl8 cell transfection and microscopy
Clone-8 (Cl8) cells were cultured as previously described (van Leeuwen et al. 1994). Cells were plated on concanavaline-coated cover glasses for 48 h and transfections using FlyFectin (OZ bioscience). After incubation for 24 h at 25 °C, the cells were fixed for 30 min with 4% paraformaldehyde and images were obtained with a Yokogawa spinning disk confocal head coupled to a Leica inverted microscope (NA 1.4, objective 100x).
In situ hybridization
Embryos were collected, dechorionated with bleach and fixed in 1 : 1 33% formaldehyde in EGTA 50 mM, and heptane for 20 min. Devitellinization was performed in 1 : 1 methanol:heptane by vortexing until the embryos fell. The latter were rinsed in PBT and in a 1 : 1 hybridization solution (Formamide 50%, SSC 2X, tRNA 1 mg/mL, Heparin 0.05 mg/mL, Roche blocking reagent 2%, Chaps 0.1%, EDTA 5 mM, Tween 20 0.1%) with PBT (PBS 1X, Tween 20 0.1%) and then incubated with the hybridization solution for 1 h at 55 °C before adding the dmlf RNA probe overnight. This probe was synthesized and labelled with the DIG RNA Labelling Kit (SP6/T7) from Roche. Embryos were washed in the hybridization solution for 1 h, incubated in a 1 : 1 hybridization solution PBT for 20 min at 55 °C and washed in PBT at room temperature. Antibody anti-digoxigenin coupled with alkaline phosphatase (DIG Nucleic acid Detection Kit, Roche) was used at 1 : 2000 for 1 h at room temperature and revelation was performed according to the manufacturer's instructions. Embryos were mounted in Spurr medium.
Immunohistochemistry
Embryos were collected, dechorionated with bleach and fixed in 1 : 1 4% formaldehyde (in PBS) and heptane for 20 min. Devitellinization was performed in 1 : 1 methanol:heptane. Embryos were blocked with 1% BSA in PBT (PBS, 0.1% Tween) for 1 h and incubated overnight with a rabbit polyclonal antidMLF antibody at 1 : 200 (Fouix et al. 2003). They were then incubated with an anti-rabbit secondary antibody coupled to Alexa 488 or Cy3 (Jackson Immunoresearch) diluted at 1 : 200 for 2 h at room temperature. For propidium iodide staining, RNaseA at 400 µg/mL was added with the secondary antibody in PBT and, after washing, embryos were incubated with propidium iodide (at 5 µg/mL, Molecular Probes) for 20 min, washed again and mounted in Citifluor (Kent).
Third instar larvae were dissected in PBS, fixed for 1 h at 4 °C in 4% paraformaldehyde-PBS and washed in PBS-0.3% Triton. The tissues were then blocked in 5% normal goat serum in PBT for 20 min and incubated with the rabbit anti-dMLF antibody (dilution 1 : 1000) at 4 °C overnight. After washing in PBT, tissues were incubated with the secondary antibody at 1 : 200 for 2 h at room temperature, washed again and mounted in Citifluor.
Samples were examined by confocal microscopy with a Leica DMR-BE microscope.
Western blot
In Fig. 1, embryos were homogenized with a piston in lysis buffer (HEPES (pH 7.5), 50 mM; NaCl, 150 mM; Igepal, 1%; EDTA, 5 mM; PMSF, 2 mM; Leupeptine, 10 µg/mL). Samples were boiled in Laemmli loading buffer (50 mM TrisCl (pH 6.8), 100 mM dithiothreitol, 2% SDS, 0.1% bromophenol blue, 10% glycerol) for SDS-PAGE electrophoresis on a 12% gel. After transfer to a nitrocellulose membrane (Schleicher & Schuell), Western blot was carried out on 20 µg of protein extract per genotype. Polyclonal rabbit anti-dMLF antibody was used at 1 : 15 000. The anti-rabbit secondary antibody coupled to the horseradish peroxydase (Vector Laboratories) was used at 1 : 10 000 and revelation was achieved using a chemiluminescence reaction (ECL Detection Reagents, Amersham). In Fig. 7, proteins were extracted from adult heads (7 days old) by the NaOH/TCA method (Riezman et al. 1983). Proteins extracts corresponding to 20 heads were fractionated on a 10% SDS-PAGE gel. A monoclonal antibody specific for tubulin (mouse, Sigma) was used to normalize the amount of proteins loaded. The secondary antibody was horseradish peroxidase-conjugated from Biovalley Vector Laboratory The enhanced chemiluminescence detection system used was the Supersignal West Pico Chemiluminescent Substrate (Pierce). For quantification, the intensity of each band (1, 2, A and B) was estimated using ImageJ and the value obtained for each dMLF variant was divided by the total amount of dMLF (sum of the values obtained for the four bands).
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Present address: aDépartement d’Hématologie, Institut Cochin, INSERM U567, CNRS UMR 8104, Université René Descartes, 75014 Paris, France;
bSchool of Biological Sciences, University of Sussex, Falmer, Brighton BN1 9QG, UK;
cDevelopmental Biology Programme, EMMeyerhofstrasse 1, 69117 Heidelberg, Germany
* Correspondence: E-mail: plessis{at}ijm.jussieu.fr |
References |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
Chen, C.H., von Kessler, D., Park, W., Wang, B., Ma, Y. & Beachy, P.A. (1999) Nuclear trafficking of Cubitus interruptus in the transcriptional regulation of Hedgehog target gene expression. Cell 98, 305–316.[CrossRef][Medline]
Cheng, S.Y. & Bishop, J.M. (2002) Suppressor of Fused represses Gli-mediated transcription by recruiting the SAP18-mSin3 corepressor complex. Proc. Natl. Acad. Sci. USA 99, 5442–5447.
Ding, Q., Fukami, S., Meng, X., Nishizaki, Y., Zhang, X., Sasaki, H., Dlugosz, A., Nakafuku, M. & Hui, C. (1999) Mouse suppressor of fused is a negative regulator of sonic hedgehog signaling and alters the subcellular distribution of Gli1. Curr. Biol. 9, 1119–1122.[CrossRef][Medline]
Dornan, D., Wertz, I., Shimizu, H., Arnott, D., Frantz, G.D., Dowd, P., O’Rourke, K., Koeppen, H. & Dixit, V.M. (2004) The ubiquitin ligase COP1 is a critical negative regulator of p53. Nature 429, 86–92.[CrossRef][Medline]
Dussillol-Godar, F., Brissard-Zahraoui, J., Limbourg-Bouchon, B., Boucher, D., Fouix, S., Lamour-Isnard, C., Plessis, A., & Busson, D. (2006) Modulation of the Suppressor of fused protein regulates the Hedgehog signaling pathway in Drosophila embryo and imaginal discs. Dev. Biol. 291, 53–66.[CrossRef][Medline]
Fouix, S., Martin-Lanneree, S., Sanial, M., Morla, L., Lamour-Isnard, C. & Plessis, A. (2003) Over-expression of a novel nuclear interactor of Suppressor of fused, the Drosophila myelodysplasia/myeloid leukaemia factor, induces abnormal morphogenesis associated with increased apoptosis and DNA synthesis. Genes Cells 8, 897–911.[Abstract]
Hart, C.M., Cuvier, O. & Laemmli, U.K. (1999) Evidence for an antagonistic relationship between the boundary element-associated factor BEAF and the transcription factor DREF. Chromosoma 108, 375–383.[CrossRef][Medline]
Hirose, F., Yamaguchi, M., Handa, H., Inomata, Y. & Matsukage, A. (1993) Novel 8-base pair sequence (Drosophila DNA replication-related element) and specific binding factor involved in the expression of Drosophila genes for DNA polymerase alpha and proliferating cell nuclear antigen. J. Biol. Chem. 268, 2092–2099.
Hirose, F., Yamaguchi, M., Kuroda, K., Omori, A., Hachiya, T., Ikeda, M., Nishimoto, Y. & Matsukage, A. (1996) Isolation and characterization of cDNA for DREF, a promoter-activating factor for Drosophila DNA replication-related genes. J. Biol. Chem. 271, 3930–3937.
Hirose, F., Yamaguchi, M. & Matsukage, A. (1999) Targeted expression of the DNA binding domain of DRE-binding factor, a Drosophila transcription factor, attenuates DNA replication of the salivary gland and eye imaginal disc. Mol. Cell. Biol. 19, 6020–6028.
Hirose, F., Ohshima, N., Shiraki, M., Inoue, Y.H., Taguchi, O., Nishi, Y., Matsukage, A. & Yamaguchi, M. (2001) Ectopic expression of DREF induces DNA synthesis, apoptosis, and unusual morphogenesis in the Drosophila eye imaginal disc: possible interaction with Polycomb and trithorax group proteins. Mol. Cell. Biol. 21, 7231–7242.
Hirose, F., Ohshima, N., Kwon, E.J., Yoshida, H. & Yamaguchi, M. (2002) Drosophila Mi-2 negatively regulates dDREF by inhibiting its DNA-binding activity. Mol. Cell. Biol. 22, 51
