Co-operation between enhancers modulates quantitative expression from the Drosophila Paramyosin/miniparamyosin gene in different muscle types

Mechanisms of Development 122 (2005) 681–694 www.elsevier.com/locate/modo Co-operation between enhancers modulates quantitative expression from the Drosophila Paramyosin/miniparamyosin gene in different muscle types Raquel Marco-Ferreres, Jorge Vivar, Juan J. Arredondo, Francisco Portillo, Margarita Cervera* Departamento de Bioquı́mica and Instituto Investigaciones Biomédicas, Facultad de Medicina, UAM-CSIC, Arzobispo Morcillo 4, 28029 Madrid, Spain Received 7 June 2004; received in revised form 10 December 2004; accepted 10 December 2004 Available online 1 January 2005 Abstract The distinct muscles of an organism accumulate different quantities of structural proteins, but always maintaining their stoichiometry. However, the mechanisms that control the levels of these proteins and that co-ordinate muscle gene expression remain to be defined. The paramyosin/miniparamyosin gene encodes two thick filament proteins transcribed from two different promoters. We have analysed the regulatory regions that control expression of this gene and that are situated in the two promoters, the 5 0 and the internal promoters, both in vivo and in silico. A distal muscle enhancer containing three conserved MEF2 motifs is essential to drive high levels of paramyosin expression in all the major embryonic, larval and adult muscles. This enhancer shares sequence motifs, as well as its structure and organisation, with at least four co-regulated muscle enhancers that direct similar patterns of expression. However, other elements located downstream of the enhancer are also required for correct gene expression. Other muscle genes with different patterns of expression, such as miniparamyosin, are regulated by other basic mechanisms. The expression of miniparamyosin is controlled by two enhancers, AB and TX, but a BF modulator is required to ensure the correct levels of expression in each particular muscle. We propose a mechanism of transcriptional regulation in which similar enhancers are responsible for the spatio-temporal expression of co-regulated genes. However, it is the interaction between enhancers which ensures that the correct amounts of protein are expressed at any particular time in a cell, adapting these levels to their specific needs. These mechanisms may not be exclusive to neural or muscle tissue and might represent a general mechanism for genes that are spatially and temporally co-regulated. q 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Enhancers; Muscle diversity; MEF2; Phylogenetic footprinting; Gene expression regulation 1. Introduction Developmental programs of muscle gene expression involve complex co-ordinated multi-step processes. These processes depend on the integration of multiple gene regulatory circuits in which modular enhancers communicate with basal promoter regions (Davidson et al., 2002; Hughes and Salinas, 1999; Kulkarni and Arnosti, 2003; McKinsey et al., 2002; Stockdale, 1997). Indeed, the creation of muscle diversity relies on three critical processes. A large group of genes encoding muscle-specific proteins must be activated in a co-ordinated manner * Corresponding author. Tel.: C34 91 497 5402; fax: C34 91 585 4401. E-mail address: margarita.cervera@uam.es (M. Cervera). 0925-4773/$ - see front matter q 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mod.2004.12.005 and high levels of protein expression achieved in a very short period of time. Furthermore, the correct stoichiometry of the distinct sets of muscle isoforms must be established, both to maintain myofibril integrity and for each particular fibre to function correctly according to their specific contractile properties (i.e.: the rate of force generation, the relaxation rate and the fatigability of the fibres; Bernstein et al., 1993; Buckingham et al., 1992; Buckingham, 1994). As the initial step in gene expression, transcription is central to these regulatory mechanisms. Moreover, the control of transcription is an elaborate process that requires specific interactions between transcription factors, cisregulatory elements in the DNA, co-factors and components of the transcriptional machinery. Indeed, the co-ordinated activity of these elements determines the local structural 682 R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 changes of the chromatin (Blackwood and Kadonaga, 1998; Davidson et al., 2002; Markstein and Levine, 2002). Recently, a regulatory mechanism has been described that controls not only the spatio-temporal expression of Troponin T and Troponin I, but also the quantities of these proteins produced in distinct Drosophila muscle types (Marin et al., 2004; Mas et al., 2004). This mechanism involves the differential yet concerted interaction of two functionally identical modular enhancers. These enhancers have been found in the first intron and in the 5 0 upstream sequence of the Drosophila TnT and TnI genes, and cooperation between these modular enhancers is essential to ensure the correct expression of these genes (Marin et al., 2004; Mas et al., 2004). By applying bioinformatic methods to analyse regulatory sequences in the Drosophila Tropomyosin gene (another component of the Troponin–Tropomyosin muscle regulatory complex), a similar modular organisation was revealed. This conservation suggests that a common mechanism might exist to establish the correct quantities of muscle proteins in each fibre-type (Mas et al., 2004). Thus, it is important to determine whether this mechanism is exclusive to TnT and TnI in vivo, or if other muscle genes share the same basic mechanism to regulate their levels of expression. Indeed, it is clear that additional studies of other muscle genes are necessary to fully explain this mechanism. In contrast to mammals, there are very few specialised muscle types generated in Drosophila and each muscle type is composed of only one type of fibre (Bate, 1990; Baylies et al., 1998; Bernstein et al., 1993). Nevertheless, flies and vertebrates mostly share evolutionary conserved molecular pathways that control muscle formation (Scott, 1994). This makes Drosophila a good model system in which to study the processes involved in the specification of muscles. We have focused our studies on the Drosophila paramyosin/ miniparamyosin gene, which encodes two invertebrate thick filament muscle proteins: Paramyosin (PM) and Miniparamyosin (mPM). These two proteins play distinct roles in different types of insect muscle, probably reflecting the particular characteristics of each muscle (Maroto et al., 1996). Paramyosin acts as a cytoplasmic protein in early embryonic development and is important for myofibril formation (Cervera et al., 2004). In contrast, miniparamyosin is important to maintain the integrity of myofibrils and for the proper function of the flight musculature (Arredondo et al., 2001b; Cervera et al., 2004). The paramyosin and miniparamyosin proteins are both encoded by the paramyosin/miniparamyosin gene, due to the activity of different promoters and alternative splicing (Fig. 1). The transcription of Paramyosin mainly involves an upstream promoter while the production of miniparamyosin uses an alternative internal promoter located in intron 7 of the gene. The two overlapping transcriptional units of the paramyosin/miniparamyosin gene act independently thus, the two promoters independently regulate the expression of the transcripts (Arredondo et al., 2001a). Whereas paramyosin is expressed in all muscles during development, miniparamyosin is found almost exclusively in the adult musculature. However, during myogenesis in the pupa, these two transcripts are expressed in the same fibers. Indeed, a transient accumulation of miniparamyosin has been detected in the head muscles of third instar larvae but it decreases rapidly during pupation (Becker et al., 1992; Maroto et al., 1996). In earlier studies, we identified several discrete regions in the respective paramyosin and miniparamyosin promoters that govern their spatio-temporal expression patterns (Arredondo et al., 2001a). Since performing a detailed characterisation of the properties of muscle-specific enhancers in transgenic organisms is a laborious process, few have been analysed in detail. Although bioinformatic methods can greatly accelerate these studies, until recently very few Fig. 1. Two independent promoters control D. melanogaster PM/mPM gene expression. Schematic representation of the gene. The exons are shown as white boxes and are specified by number, except exon 1b which encodes the 5 0 end of miniparamyosin and is in grey. The promoter regions controlling the two transcriptional units are shown. DME, distal muscle enhancer. AB, BF and TX are the conserved elements involved in the regulation of miniparamyosin expression. R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 of the predictions have been tested in multi-cellular animals (Frith et al., 2001; Pennachhio and Rubin, 2001; Sandelin and Wasserman, 2004). Here, we have analysed the regulation of the PM/mPM gene by linking a reporter gene to distinct genomic sequences situated in the two promoters controlling this gene and analysing reporter gene expression in vivo in germ line transformants of Drosophila. We show that the complex spatio-temporal pattern of Drosophila paramyosin and miniparamyosin expression depends on several developmentally regulated transcriptional enhancers in the gene locus. Co-operation between these enhancers is essential to produce the correct quantities of protein in each muscle type, and to ensure adequate gene expression. We demonstrate that the levels of expression in each particular muscle are established through concerted interactions between these elements. Moreover, the basic regulatory design described here, appears to be conserved in other Drosophila muscle genes. 2. Results 2.1. Paramyosin expression is promoted by a distal muscle enhancer in all muscle types Bioinformatic tools are tremendously useful when attempting to define elements that play a role in transcriptional control. Indeed, accurate and precise sequences have been identified using this approach (Konig et al., 2002). We have already identified several positive regulatory elements that govern the spatial and temporal patterns of paramyosin expression, all in the 5 0 upstream sequence of the gene (Arredondo et al., 2001a). Indeed, no reporter activity was detected when intron 1 alone was analysed in gene fusion assays (data not shown). To better define these regulatory elements, sequences from D. melanogaster and D. pseudoobscura species were aligned enabling us to verify the striking conservation of sequences upstream of the previously identified MEF-E region (Fig. 2A and Arredondo et al., 2001a). As a result, we performed classical gene fusion assays, using the newly identified conserved region between K2.3 and K1.4 kb to direct the expression of a lacZ reporter gene (Fig. 2B). This element was compared with the previously identified MEF-E region (located between K1.7 and K1.4 kb) that contains a functional MEF2 site and three E boxes (Arredondo et al., 2001a). In these assays, both fragments were linked to a heterologous hsp70promoter in phase with the reporter gene. The two constructs will be referred to as (K2.3/K1.4)PM and (K1.7/K1.4)PM respectively, thereby identifying the upstream position of the fragments with respect to the transcription start site. As a control, the PM 4 (K4.0/C1) transgenic line containing the entire 4 kb upstream of the transcription start site of the gene was also analysed (Arredondo et al., 2001a). 683 When the levels of reporter gene expression were tested (Fig. 2B and Table 1), the entire region situated between K1.4 and K2.3 acted as a very strong enhancer, directing high levels of expression in all the major muscle groups of the embryo and larva. High levels of expression were also seen in all adult muscles except for the specialised flight muscles. We referred to this region as the Distal Muscle Enhancer (DME) element. In both, the (K1.7/K1.4)PM and PM4 transgenic lines, lower levels of reporter gene activity were observed in all adult and larval muscle types (Fig. 2B; Arredondo et al., 2001a). Moreover, (K1.7/K1.4)PM reporter gene expression was relatively irregular in the larval fibres indicating a deregulation of gene expression at this stage (upper right panel in Fig. 2B). Our in silico analysis revealed that this distal muscle enhancer (DME) contains several putative binding sites for myogenic factors. As in other genes encoding contractile proteins, we confirmed that the MEF2 sites were important for the control of paramyosin gene expression (Black and Olson, 1998). This analysis revealed that the three MEF2 sites, the two E boxes and one CF2 binding site were conserved in the orthologous D. pseudoobscura gene (Fig. 2A). The conserved spatial arrangement of these binding sequences for myogenic factors within larger conserved domains of 30–80/90 bp was evident in annotated maps. In these maps, the degree of sequence conservation and the precise location of the matches were established (Fig. 2A). A similar spatial organisation has also been described in other co-regulated muscle genes (TnT, TnI and Tm) indicating the importance of the number and arrangement of the binding sites for myogenic factors in this type of enhancer elements (Marin et al., 2004; Mas et al., 2004). Several CF2 and PDP1 binding sites were identified in the sequences situated downstream from the DME enhancer although, phylogenetic footprinting revealed that only one of the CF2 sites was conserved in the orthologous D. pseudoobscura gene. Both CF2 and PDP1 have previously been shown to play important roles in muscle development (Bagni et al., 2002; Reddy et al., 2000). Other unknown sequences of 20–30/40 bp also maintained a high degree of sequence conservation and precisely located matches were identified in this region (Fig. 3A). Thus, we decided to study the region in between K1.4 and K0.76 as a candidate modulator of the intense distal enhancer activity. We subdivided this region into two smaller fragments, maintaining the MEF-E region in both the constructs, and tested their activity in all tissues of the transgenic lines at different developmental stages (Fig. 3B). The resulting expression patterns indicated that downstream sequences flanking the MEF-E region clearly reduced transcription in larval and adult muscles (summarised in Table 1). Based on the high degree of sequence conservation in this region between distinct Drosophila species, one possible explanation for this result might be the presence of a negative modulator situated in the region between K1.4 and K0.76 (Fig. 3A). However, we cannot rule out the possibility that 684 R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 Fig. 2. A distal muscle enhancer governs strong paramyosin expression in all muscle types except for the indirect flight muscles. In (A) the alignment of the Distal Muscle Enhancer sequence situated between K2295 and K1387 of the D. melanogaster PM/mPM gene and the related sequences in D. pseudoobscura is shown. The conserved binding sites for myogenic motifs, including the MEF2, E box and CF2 are indicated. The vertical bar indicates the 5 0 end of the (K1.7/K1.4)PM construct. In addition, the MEF-E region is shown in grey. The sequences situated upstream of the DME are not conserved. In (B) the b-galactosidase staining of 3rd instar larvae and thin sections of thorax and abdomen of 1-day-old flies transformed with the PM4, (K2.3/K1.4)PM or (K1.7/K1.4)PM construct are shown. In the upper part of the photographs, there is a schematic representation of the constructs analysed in the pH-Pelican 685 R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 Table 1 b-Galactosidase reporter expression driven by sequences upstream of the PM transcription unit b-galactosidase expression Analysed lines TDT IFM Leg muscles DFM Abdominal hypodermic muscles Visceral muscles Larval muscles 2 CCC K CCC CCC CCC CCC 3 C K CC CC CC CC 2 CCCC K CCCC CCCC CCCC CCCC 5 C K CC CCC CCC CC/K 4 K K C/K CC CC C 5 K K CC C C K Comparison of b-galactosidase reporter expression driven by the fragments represented at the left. The relative levels of expression are compared to the levels in (K2.3/K1.4) lines, the highest expressing lines. this effect was partially caused by an increase in the distance between the enhancer and the basal promoter. 2.2. Miniparamyosin expression in all muscle types depends on the co-operation of two muscle enhancers, AB and TX, and a modulator, BF Previous studies on miniparamyosin gene expression identified three regulatory elements that directed miniparamyosin expression in the context of the complex endogenous promoter, the AB, TX and BF elements (Arredondo et al., 2001a). In contrast to the majority of cis-regulatory elements that control muscle protein expression, no myogenic binding sites have been found in these. Indeed, the functional complexity introduced by the co-ordinated activity of these elements complicates the interpretation of the results. Hence, we attempted to define the exact function of each of these elements in the miniparamyosin cis-regulatory region, using classical gene fusion assays. The AB, TX and BF fragments were individually linked to the heterologous hsp70 promoter to direct the expression of the b-galactosidase/reporter gene (see Section 4), and b-galactosidase activity was assayed in the transgenic lines generated (Fig. 4 and Table 2). Lines that carried intron 7 along with the three elements (mP1.7) reproduced the endogenous expression of miniparamyosin and were used as controls (Arredondo et al., 2001a). The AB element independently acts as a specific and very strong enhancer element, directing very high levels of expression in the specialised indirect flight muscles (IFM). Indeed, this element acts exclusively in these muscles although non-specific activity was also observed in larvae and in the abdomen oenocytes (Fig. 4 and Table 2). On the other hand, the TX element governs high levels of reporter expression in the tergal depressor of the throchanter muscle (TDT), the direct flight muscles (DFM), the leg and visceral muscles, and in the larval head muscles. This element also induces lower levels of expression in the hypodermic abdominal muscles and the IFM. Interestingly, the patterns of muscle expression driven by AB and TX were practically complementary. Curiously, no activity was observed in any of the transgenic lines carrying the conserved BF element (Fig. 4 and Table 2). When transgene expression was driven by a construct harbouring both AB and TX, all larval and adult muscles were intensely stained. The presence of these two elements in the transgenic lines resulted in spatio-temporal patterns and levels of reporter expression equivalent to the addition of that obtained in the lines carrying either AB or TX alone (Fig. 4 and Table 2). Indeed, the expression of the hypodermic abdominal muscles was much higher than the simple addition of that driven by the two elements alone. In summary, the levels of reporter expression in these lines were higher than in the control lines (mP1.7). Hence the possibility arose that the conserved BF element might modulate these levels of expression. To test whether the conserved BF element can indeed influence the expression driven by AB and TX, we combined these elements in new lines, respecting the relative position of the elements in the endogenous gene (Fig. 5 and Table 2). 3 bgal plasmid in which the MEF2, PDP1 and CF2 binding sites, and the E boxes are represented. One representative line of the two analysed is presented in each case. The measurement of b-galactosidase activity was performed after long reaction times (16 h) in the PM4 and (K1.7/K1.4)PM lines and, due to the strong b-galactosidase activity, after short reaction times (30 min) in (K2.3/K1.4)PM lines. In larvae and adult sections anterior is left and posterior is right. IFM, indirect flight muscles (/); TDT, tergal depressor of the trochanter muscle (*); abdominal hypodermic muscles ( ). 686 R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 Fig. 3. The region situated between K1.4 and K0.76, causes a clear reduction of transcription in larval and adult muscles. In (A), the alignment of the sequence situated between nucleotides K1384 and K764 of the D.melanogaster PM/mPM gene and the related sequences in D. pseudoobscura is shown. The vertical bar at nucleotide K1156 indicates the 3 0 end of the fragment cloned in the (K1.7/K1.15)PM construct. (B), b-galactosidase staining of 3rd instar larvae and thin sections of the thorax and abdomen of one day old flies transformed with the (K1.7/K1.4)PM, (K1.7/K1.15)PM and (K1.7/K0.76)PM constructs is shown. In the upper part of the photographs, we have included a schematic representation of the constructs in the pH-Pelican bgal plasmid. In the constructs, the MEF2, PDP1, CF2 binding sites and the E boxes are represented. One representative line of the two analysed is presented in each case, the staining patterns being visualised after 16 h. In the larval and adult sections, anterior is left and posterior is right. IFM, indirect flight muscles (/); TDT, tergal depressor of the trochanter muscle (*); abdominal hypodermic muscles ( ). The presence of the BF element produced an important decrease in the levels of reporter gene expression in the indirect flight muscles when compared with flies harbouring the AB element alone. The levels of reporter expression in the IFM were similar with the AB/BF construct to those in control lines (mP1.7). In contrast, the flies harbouring both the TX and BF elements, showed a slight decrease in the levels of expression in the TDT. Interestingly, these results suggest that the interaction of the BF element with the AB element acts as a negative modulator in terms of the quantity R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 687 Fig. 4. AB and TX elements are involved in controlling miniparamyosin expression in distinct muscle types. b-galactosidase staining of microdissected 3rd instar larvae, abdomens and thin sections of the thorax of newly emerged flies transformed with distinct constructs containing the AB element (AB), BF element (BF), TX element (TX), a combination of both AB and TX (AB–TX) or intron 7 containing all three elements (mP 1.7), linked to the heterologous hsp70 promoter and b-galactosidase gene. The reaction time for b-galactosidase staining was 1 h. The strong staining in the oenocytes (oe) in larva and abdomens in the AB lines, as well as of the pericardial cells (pc) in the abdomens, is non-specific. In the adult thoraxic sections and dissected abdomens anterior is up; in the larvae, anterior is to the left. IFM, indirect flight muscles (/); TDT, tergal depressor of the trochanter muscle (*); hdm, hypodermic dorsal muscles; hvm, hypodermic ventral muscles. of protein produced in the IFM. However, this effect was not clear in the TDT when BF was coupled to the TX element. Moreover, the AB and TX elements appeared to co-operate to increase the levels of expression in the hypodermic abdominal muscles. 2.3. The transcription factor PDP1 interacts with the BF element Most of the known myogenic factors are involved in the formation of the embryonic, and larval, as well as the adult musculature (McKinsey et al., 2002; Taylor, 1998, 2002). Factors that exclusively specify the adult musculature have not been reported. Accordingly, our in silico analysis failed to reveal any known myogenic factor binding sites in the three elements of the miniparamyosin promoter in the orthologous D. pseudoobscura gene. In order to clarify this issue, we decided to perform a screen using the one hybrid assay to identify transcription factors that bind specifically to the BF element. The size of this element facilitates the use of this assay, in which five tandem copies of BF (34 bp) were used to direct the yeast 688 R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 Table 2 b-Galactosidase reporter expression driven by selected fragments within the miniparamyosin promoter b-galactosidase expression Analysed lines TDT IFM Leg muscles DFM Abdominal hypodermic muscles 2 CCC C CC CC C 4 K CCC K K K 2 K K K K K 2 CCCC C/K CCC C CCC 3 CCCC CCC CCC CC CCC 4 K C K K K 4 K CCC K K K 3 CCC K CCC C CCC 3 CCCC C/K CCC C CCC Larval muscles Comparison of b-galactosidase expression driven by the AB, TX, BF, AB–BF, AB–TX and TX–BF fragments. b-galactosidase activity in mP1.7 muscles was used as a reference. The level of intensity of the other lines was determined by comparison with these lines. HIS3 gene. We screened a D. melanogaster adult cDNA library carrying the GAL4 activation domain (see Section 4), and identified 273 clones that were sequenced and analysed. A BLAST analysis revealed that 60% of the clones identified corresponded to the alfa, beta and delta isoforms of PDP1 (Reddy et al., 2000). A closer analysis of the BF element revealed the presence of the sequence AAAATTG at its 5 0 end, which is very similar to the PDP1 consensus sequence, TAAAACATTG (Lin et al., 1997b and Fig. 6A). The remaining clones appeared to be non-specific and mainly corresponded to mitochondrial cDNAs. The Pdp1 gene is a very complex gene, encoding at least six isoforms that are differentially expressed in distinct tissues of the Drosophila embryo. Most of the clones we isolated corresponded to the a and b PDP1 isoforms that are specifically expressed in pre-somatic and somatic muscles, and in the hindgut (Reddy et al., 2000). We therefore confirmed that the Drosophila PDP1 protein can bind to the BF sequence through band shift assays and that an oligonucleotide containing the mutated PDP1 site was unable to compete the protein binding (Fig. 6B). Then, we decided to analyse the role of PDP1 in miniparamyosin gene regulation and several BF–TX (DPDP1) and AB–BF (DPDP1) transgenic lines were generated in which the PDP1 site in the BF element was deleted, to determine how critical the interaction of PDP1 with this sequence was (Table 2 and Fig. 5). The absence of the PDP1 site in the BF-TX (DPDP1) lines produced an increase in reporter expression in adult muscles and in the head muscles of third instar larvae (Table 2 and Fig. 5). The expression pattern of the transgene was respected and was similar to that observed in the TX lines (Fig. 5). A similar effect was seen in the AB–BF (DPDP1) transgenic lines where the expression pattern is equivalent to that of the AB lines (Table 2 and Fig. 5). The effect was observed in all the lines analysed. The expression of the reporter gene in adult and larval muscles clearly indicates that the binding of PDP1 to the BF sequence plays an important role in modulating the activity of AB and TX (Figs. 5). The failure of PDP1 to bind to this sequence seems to abolish the influence of BF on the expression of miniparamyosin. Thus, the BF fragment does, indeed, modulate AB and TX enhancer activity through its association with PDP1. Since PDP1 is principally thought to be an activator and a and b isoforms are expressed in muscle tissue at the same stages of development as mPM (data non shown), we expected to observe a decrease of reporter activity in the muscles. Instead, the absence of the PDP1 binding site in the BF element produces an increase in reporter activity. However, the result was not completely surprising, since PDP1-delta isoform functions as a dominant negative inhibitor of transcription forming heterodimers with other PDP1 isoforms (Reddy et al., 2000; Lin et al., 1997b). We believe the modulator complexes that bind BF element may contain distinct PDP1 isoforms and that the composition of these complexes might vary during development. This raises the possibility that these complexes will act as activators or inhibitors, depending on their composition and the stage at which they act. In fact, these complexes are likely to be made up of transcription factors, cofactors and other components, some of which will be common to different or all complexes and some of which will vary. The elimination of one factor in the complex will influence other components creating distinct situations. 3. Discussion The functional complexity produced by the action of many cis-regulatory elements makes it difficult to study the true activity of these elements. However, while laborious, the detailed functional dissection of these cis-regulatory R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 689 Fig. 5. The BF element negatively modulates the expression levels produced by the AB and TX elements. b-galactosidase staining of 3rd instar larvae and thin sections of the thorax and abdomen of newly emerged flies transformed with distinct constructs containing distinct combinations of BF or BF (DPDP1) with the AB and TX elements (AB–BF and BF–TX). In parallel and as a control, the mP1.7 lines were also stained for b-galactosidase activity. The reaction time for b-galactosidase staining was 30 min. In the adult thoraxic sections and dissected abdomens anterior is up; in the larvae, anterior is to the left. IFM, indirect flight muscles (/); TDT, tergal depressor of the trochanter muscle (*); abdominal hypodermic muscles ( ). regions helps us to gain a more complete picture of how they might interact. Our in vivo and in silico studies on the PM and mPM promoter regions have identified independent discrete regulatory regions and some of their functional relationships. Indeed, in this way we have identified a 900 bp distal muscle enhancer in the paramyosin regulatory region. The very strong activation of reporter expression by this enhancer in virtually all Drosophila muscles is regulated by downstream sequences, situated between K1.4 and 0.76. These downstream sequences negatively modulate the levels of expression driven by the DME enhancer in larval and adult muscles (Fig. 3). 690 R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 Fig. 6. The transcription factor PDP1 binds to the BF sequences. In (A) the BF sequence and the binding site for PDP1 protein. The BFExtmt sequence corresponds to the BF element containing a mutated PDP1 binding site. In (B) electrophoretic mobility shift assays show that in vitro transcribed and translated PDP1 protein specifically bound to the sequence found in the BF element in the miniparamyosin promoter (lane 2). Competition for PDP1 binding to this site can be seen with the wild type BF oligonucleotide sequence (lanes 3 and 4), but not with a mutant BFExtmt oligonucleotide sequence (lanes 5 and 6). Addition of different amounts of PDP1 antibody produces a super-shift of the bound complex (lanes 7–9) while addition of pre-immune serum failed to induce a supershift of the bound complex (lanes 10 and 11). The structure and organisation of this paramyosin DME enhancer reveals certain similarities to the Troponin T and Troponin I enhancer elements (Fig. 7; Marin et al., 2004; Mas et al., 2004). This is also reflected by the fact that the three genes have the same pattern of expression during development, and that they are present in all muscles at the two distinct developmental stages in Drosophila. These enhancers are approximately 800–900 bp in size and they direct expression in embryonic and larval muscles, as well as in adult muscles. Interestingly, within these enhancer sequences, the three MEF2 binding sites that are conserved in D. pseudoobscura are contained within larger conserved regions of between 35 and 80 bp. Moreover, in vitro transcribed and translated MEF2 binds specifically to these sequences (data not shown), and we have demonstrated previously that at least one of the MEF2 binding sites is functionally relevant in vivo (Arredondo et al., 2001a). Furthermore, at least one conserved CF2 binding site is also contained in these enhancers. It also appears that similar muscle enhancers may regulate the expression of the Drosophila Tropomyosin gene (Fig. 7; Gremke et al., 1993; Lin et al., 1996; Lin and Storti, 1997). The fact that these features are so widely conserved suggests that this general structure might be important for the co-ordinated regulation of this group of developmental regulated muscle genes. Recently, Erives and Levine (2004) reached the same conclusion by examining the organisation of four co-ordinately regulated neurogenic enhancers that direct similar patterns of gene expression in the early Drosophila embryo (Erives and Levine, 2004). A subset of putative regulatory elements found within the four neurogenic enhancers are arranged similarly. Through these shared features, they identified a new neurogenic enhancer in the distantly related Anopheles genome. We have found two muscle enhancers in the miniparamyosin promoter region, AB and TX, which independently drive expression in distinct Drosophila muscles. The AB element drives high levels of reporter expression exclusively in the specialised indirect flight muscles, whereas the TX element does so in the rest of the Drosophila musculature that expresses miniparamyosin (Fig. 4). These two elements, AB and TX, are responsible for the spatio-temporal expression of miniparamyosin but interact with a third element situated in between them, BF, to modulate the expression levels produced in each particular muscle. Thus, reporter activity was diminished in the IFM of AB–BF transgenic flies when compared with the activity induced by AB alone. Although a milder decrease in activity in the TDT R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 691 Fig. 7. Phylogenetically conserved domains within the cis-regulatory regions of the Drosophila PM/mPM, TnT, TnI and Tm2 genes. A schematic representation of conserved binding sites for myogenic motifs, including MEF2, PDP1 and CF2, and the E-boxes is presented. Several conserved sequences that are not targets for known myogenic factors were found (grey squares). was observed when BF was associated with TX. In adult hypodermic muscles this situation changes as all three elements interact to produce the correct protein levels. Thus, TX produces low levels of reporter activity in hypodermic muscles whereas the AB element remains inactive in these muscles. When hypodermic muscles from mP1.7 flies that contain all three elements are examined, protein expression levels are higher and correspond to the wild type endogenous miniparamyosin protein (Arredondo et al., 2001a). When other muscle enhancers in Drosophila are considered (Fig. 7), it appears that the structure and organisation of the three regulatory elements controlling miniparamyosin expression is completely distinct. First of all, these elements are smaller, ranging from 260 bp for the TX element to 34 bp for the BF element. Secondly, none of them contain MEF2 binding sites, even though MEF2 is thought to be essential for muscle protein expression (Black and Olson, 1998). Indeed, the apparently functional PDP1 binding site in the BF modulator element seems to be the only myogenic specific binding site that can be found. The PDP1 transcription factor plays a critical role in the BF element, since the absence of its binding site seems to eliminate the effect of the BF element. However, phylogenetic footprinting analysis reveals a striking conservation of the three elements in the D. pseudobscura and D. virilis orthologous regulatory regions, highlighting the importance of these elements for miniparamyosin expression. It is important to note that miniparamyosin is principally distinguished from paramyosin (or other contractile muscle proteins) in that it is almost exclusively expressed in the adult musculature or in the head muscles of third instar larvae (Becker et al., 1992; Maroto et al., 1996). The other muscle genes are expressed at all stages of Drosophila development, in late embryos, larvae, late pupae and adults. Previous studies from our laboratory revealed that the correct levels of Troponin T protein in each Drosophila muscle-type are established through the direct or indirect interactions between two elements: the Upstream Regulatory Element (URE) and the Intronic Regulatory Element (IRE; Mas et al., 2004). These two elements can independently activate TnT transcription in all muscle types. However, in the case of the TnT, we demonstrated that the two elements collaborate to ensure the full recapitulation of endogenous TnT gene expression. Moreover, the contribution of each regulatory element in achieving correct TnT expression varies between muscle types. The synergism of the URE and IRE in larval muscles was higher than that seen in adult muscles. Interestingly a similar mechanism has been described in the Drosophila TnI gene (Marin et al., 2004), which is also controlled by two enhancer elements, located upstream and downstream of the start site of the gene. 692 R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 It is clear that certain similarities in cis-transcription are observed when comparing the molecular mechanisms that drive the expression of Drosophila muscle genes with the same developmental expression patterns. TnT and TnI genes have two enhancer regulatory elements, URE and IRE, that differentially interact to ensure the correct levels of expression in each particular muscle. The data presented here indicate that this basic mechanism is not exclusive to these genes, but that it is also relevant to the other muscle specific paramyosin/miniparamyosin and tropomyosin genes. Moreover, chromatin remodelling must be also incorporated into this puzzle. Although the precise mechanism is still unknown, it is clear that this organism has developed two distinct levels of transcriptional regulation that must be integrated to ensure both the proper temporal and tissue specific expression, and the adequate quantity of protein in each particular muscle. In the paramyosin/miniparamyosin gene, as in the TnT and TnI genes, several enhancers within the gene locus are responsible to ensure the specific spatio-temporal expression pattern, and through co-operative interactions, the levels of expression in each muscle-type. However, the basic regulatory design described here, appears to be different for muscle genes with other expression patterns, such as miniparamyosin, actin 88F or flightin. Miniparamyosin is expressed in all Drosophila adult muscles whereas flightin and actin 88F are almost exclusively expressed in indirect flight muscles (Maroto et al., 1996; Nongthomba et al., 2001; Vigoreaux et al., 1993). In the regulatory regions of the flightin and actin 88F genes the search for myogenic binding sites, distal muscle-like enhancers, or AB-like elements has been unsuccessful. In the miniparamyosin enhancer only the latter was found, this being specific for IFM expression. This is not surprising, since although all these genes are only expressed in the adult, they are expressed in distinct muscle subsets. Based on the data presented above and that of others (Erives and Levine, 2004; Marin et al., 2004; Mas et al., 2004), we conclude that co-regulated genes with the same expression pattern during development have specific enhancers with a similar structure and organisation that are responsible for their strong specific tissue-activation. These enhancer activities seem to be modulated through interactions between themselves and with other regulatory elements. Interactions between the regulatory elements are necessary to ensure that the appropriate amounts of protein are established in each particular muscle. Hence, we propose a mechanism of transcriptional regulation in which similar enhancers in different genes are responsible for the spatio-temporal specificity of gene expression, and wherein the interaction between enhancers ensures that the correct amounts of protein are expressed at any particular time in a cell. These mechanisms might not be exclusive to neural or muscle tissues but may represent a general mechanism for those genes with the same spatio-temporal co-regulation. 4. Experimental procedures 4.1. Plasmid vectors, P-transformation and Drosophila transgenic lines The D. melanogaster genomic clones containing the distinct sequences upstream of the transcription initiation sites of the paramyosin and miniparamyosin were subcloned and sequenced as described previously (Arredondo et al., 2001a). Distinct fragments of these regions were cloned into P-transformation vectors, respecting their native orientation relative to the basal promoters. In BF–TX (DPDP1) and AB–BF (DPDP1) lines, PDP1 site was deleted. The main transcription initiation starting point of the gene in the clones identified is referred to as C1 bp in our map (Arredondo et al., 2001a). All fragments were cloned into pH-Pelican vector (Barolo et al., 2000) whose flanking sequences are of the insulator type to prevent local enhancer effects on reporter expression. These constructs were sequenced, and one copy of each was inserted into the lines analysed. Two transformant lines were analysed for each construct. The generation of germ line transgenic flies using the P element-mediated transformation technique was essentially as described (Spradling and Rubin, 1982). 4.2. Histochemistry Standard procedures were used to assay b-galactosidase enzyme activity in larvae and adults of transgenic lines, with only minor modifications (Arredondo et al., 2001b; Sullivan et al., 2000). At least two different lines with each construct were analysed. Microdissected third instar larvae and adult cryostat sections were fixed and stained as previously described (Meredith and Storti, 1993). The activity of b-galactosidase was used as a means to compare the transcriptional efficiency between different constructs. In larval and adult muscles, a rough quantification of the expression levels between different constructs was achieved by visually monitoring the time required for the blue reaction products from the X-gal substrate/b-galactosidase reaction to appear. The complete expression patterns were visualised after two hours of staining in larvae and microdissected abdomen, and after 60 min in thoracic sections if not otherwise specified. 4.3. Yeast one-hybrid screen The MATCHMAKER One-Hybrid System (Clontech) was used according to the manufacturers’ instructions. Briefly, five tandem repeats of the BF sequence (5 0 -AGA AAA TTG CGC AGT TGC TGT GAT TAC ACA AT-3 0 ) were inserted upstream of the HIS3 reporter gene in the pHISi (Clontech) plasmid, using the appropriate restriction sites. The inserts were then sequenced to confirm the presence and the integrity of the repeats. BF reporter constructs were linearised with NcoI and integrated into R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 the genome of the S cerevisiae strain YM4271 to generate a BF reporter strain. Background HIS3 activity was ablated using 45 mM 3-amino-1,2,4-trizole (3-AT; Sigma). Screening BF reporter strains with a Drosophila adult cDNA/GAL4 activation domain fusion library led to the identification of genes encoding proteins that bound to the BF element. Plasmids from positive clones were isolated from the yeast and transformed into Escherichia coli DH5a. Each cDNA insert was sequenced and compared to known sequences in the GenBank database. To discard false positives, we transformed the isolated clones into two reporter strains with tandem repeats of the mutated BF sequence. 4.4. In vitro translation and EMSA The PDP1/pSPUTK plasmid (1 mg) was incubated with the SP6 TNT wheat germ extract coupled transcription/ translation system for 120 min at 30 8C (Promega). Protein production was confirmed by Western blot and the in vitro synthesised product was then subjected to EMSA. A 1 ml aliquot of the in vitro-translated product was incubated for 30 min with 0.1 ng of a 32P-labelled oligonucleotide containing the BF sequence and 1 mg of poly (dI–dC) in binding buffer (60 mM KCl, 2 mM MgCl2, 10% glycerol, 0.1 mM EDTA, 1 mM DTT, 20 mM HEPES [pH 7.5]). For competitive analysis, a 100-fold or 300-fold molar excess of competitor DNA’s were added to the binding mixture. Supershift assays were carried out by incubating with an anti-PDP1 serum for 10 min at room temperature, or with a preimmune serum, before adding the probe. Once binding had been completed, the mixture was separated by electrophoresis on a 5% polyacrylamide gel in 0.5! TBE, which were then dried and exposed to autoradiographic film at K70 8C for 16 h. The sequence of the wild type and mutant BF oligonucleotides is available upon request. 4.5. Genomic analyses Sequence comparison between different Drosophila species, and searches for transcription factor binding sites in short genomic sequences near the transcription initiation site of the genes, were performed with the Gene Jockey II program (Biosoft, Cambridge). The genomic D. pseudoobscura sequence database was accessed through the Drosophila Genome Project (http://www.hgsc.bcm.tmc.edu/ projects/drosophila). The consensus sequences used to search for binding sites were: YTAWWWWTAR for MEF2 (Taylor et al., 1995); RTATATRTA for CF2 (Gogos et al., 1996); and RTTTWAYGTAAY for PDP1 (Reddy et al., 2000). The genomic sequences around the paramyosin gene and the relative positions of the distinct genes were obtained from the GenBank at NCBI and FlyBase (http://flybase.bio.indiana.edu). 693 4.6. Nucleotide sequence accession numbers The complete upstream and downstream sequences of the paramyosin/miniparamyosin gene from D. melanogaster were submitted to GenBankTM/EMBL Data Bank with the accession numbers AJ243067 and AJ243068, respectively. Acknowledgements This research was supported by a grant BMC2001-1454 from the DGICYT (Spanish Ministry of Education, Culture and Sports). R. Marco Ferreres was a pre-doctoral fellow of the Spanish Ministry of Education, Culture and Sports and J. Vivar is a pre-doctoral fellow at the Universidad Autónoma de Madrid. We are grateful to Dr R. Storti and Dr H. Nguyen for the generous gifts of plasmids and antibodies MEF2 and PDP1, respectively. We thank Drs A. Cano and R. Garesse for their critical reading of the manuscript, and we are also grateful to Vanesa Santos for her expert technical assistance. References Arredondo, J.J., Marco-Ferreres, R., Maroto, M., Cripps, R.M., Marco, R., Bernstein, S.I., Cervera, M., 2001a. Control of Drosophila paramyosin/miniparamyosin gene expression. Differential regulatory mechanisms for muscle-specific transcription. J. Biol. Chem. 276, 8278–8287. Arredondo, J.J., Mardahl-Dumesnil, M., Cripps, R.M., Cervera, M., Bernstein, S.I., 2001b. Overexpression of miniparamyosin causes muscle dysfunction and age-dependant myofibril degeneration in the indirect flight muscles of Drosophila melanogaster. J. Muscle Res. Cell Motil. 22, 287–299. Bagni, C., Bray, S., Gogos, J.A., Kafatos, F.C., Hsu, T., 2002. The Drosophila zinc finger transcription factor CF2 is a myogenic marker downstream of MEF2 during muscle development. Mech. Dev. 117, 265–268. Barolo, S., Carver, L.A., Posakony, J.W., 2000. GFP and beta-galactosidase transformation vectors for promoter/enhancer analysis in Drosophila. Biotechniques 29, 726, 728, 730, 732. Bate, M., 1990. The embryonic development of larval muscles in Drosophila. Development 110, 791–804. Baylies, M.K., Bate, M., Ruiz Gomez, M., 1998. Myogenesis: a view from Drosophila. Cell 93, 921–927. Becker, K.D., O’Donnell, P.T., Heitz, J.M., Vito, M., Bernstein, S.I., 1992. Analysis of Drosophila paramyosin: identification of a novel isoform which is restricted to a subset of adult muscles. J. Cell Biol. 116, 669–681. Bernstein, S.I., O’Donnell, P.T., Cripps, R.M., 1993. Molecular genetic analysis of muscle development, structure, and function in Drosophila. Int. Rev. Cytol. 143, 63–152. Black, B.L., Olson, E.N., 1998. Transcriptional control of muscle development by myocyte enhancer factor-2 (MEF2) proteins. Annu. Rev. Cell Dev. Biol. 14, 167–196. Blackwood, E.M., Kadonaga, J.T., 1998. Going the distance: a current view of enhancer action. Science 281, 61–63. Buckingham, M.E., 1994. Muscle: the regulation of myogenesis. Curr. Opin. Genet. Dev. 4, 745–751. 694 R. Marco-Ferreres et al. / Mechanisms of Development 122 (2005) 681–694 Buckingham, M., Houzelstein, D., Lyons, G., Ontell, M., Ott, M.O., Sassoon, D., 1992. Expression of muscle genes in the mouse embryo. Symp. Soc. Exp. Biol. 46, 203–217. Cervera, M., Arredondo, J.J., Marco-Ferreres, R., 2004. Paramyosin and Miniparamyosin, in: Vigoreaux, J.O. (Ed.), Nature’s Versatile Engine: Insect Flight Muscle Inside and Out. Landers Bioscience Publishers, Georgetown, TX. Davidson, E.H., Rast, J.P., Oliveri, P., Ransick, A., Calestani, C., Yuh, C.H., et al., 2002. A genomic regulatory network for development. Science 295, 1669–1678. Erives, A., Levine, M., 2004. Coordinate enhancers share common organizational features in the Drosophila genome. Proc. Natl. Acad. Sci. USA 101, 3851–3856. Frith, M.C., Hansen, U., Weng, Z., 2001. Detection of cis-element clusters in higher eukaryotic DNA. Bioinformatics 17, 878–889. Gogos, J.A., Jin, J., Wan, H., Kokkinidis, M., Kafatos, F.C., 1996. Recognition of diverse sequences by class I zinc fingers: asymmetries and indirect effects on specificity in the interaction between CF2II and ACT-rich elements. Proc. Natl. Acad. Sci. USA 93, 2159–2164. Gremke, L., Lord, P.C., Sabacan, L., Lin, S.C., Wohlwill, A., Storti, R.V., 1993. Coordinate regulation of Drosophila tropomyosin gene expression is controlled by multiple muscle-type-specific positive and negative enhancer elements. Dev. Biol. 159, 513–527. Hughes, S.M., Salinas, P.C., 1999. Control of muscle fibre and motoneuron diversification. Curr. Opin. Neurobiol. 9, 54–64. Konig, S., Burkman, J., Fitzgerald, J., Mitchell, M., Su, L., Stedman, H., 2002. Modular organization of phylogenetically conserved domains controlling developmental regulation of the human skeletal myosin heavy chain gene family. J. Biol. Chem. 277, 27593–27605. Kulkarni, M.M., Arnosti, D.N., 2003. Information display by transcriptional enhancers. Development 130, 6569–6575. Lin, S.C., Storti, R.V., 1997. Developmental regulation of the Drosophila Tropomyosin I (TmI) gene is controlled by a muscle activator enhancer region that contains multiple cis-elements and binding sites for multiple proteins. Dev. Genet. 20, 297–306. Lin, M.H., Nguyen, H.T., Dybala, C., Storti, R.V., 1996. Myocyte-specific enhancer factor 2 acts cooperatively with a muscle activator region to regulate Drosophila tropomyosin gene muscle expression. Proc. Natl. Acad. Sci. USA 93, 4623–4628. Lin, S.C., Lin, M.H., Horvath, P., Reddy, K.L., Storti, R.V., 1997. PDP1, a novel Drosophila PAR domain bZIP transcription factor expressed in developing mesoderm, endoderm and ectoderm, is a transcriptional regulator of somatic muscle genes. Development 124, 4685–4696. Marin, M.C., Rodriguez, J.R., Ferrus, A., 2004. Transcription of Drosophila troponin I gene is regulated by two conserved, functionally identical, synergistic elements. Mol. Biol. Cell 15, 1185–1196. Markstein, M., Levine, M., 2002. Decoding cis-regulatory DNAs in the Drosophila genome. Curr. Opin. Genet. Dev. 12, 601–606. Maroto, M., Arredondo, J., Goulding, D., Marco, R., Bullard, B., Cervera, M., 1996. Drosophila paramyosin/miniparamyosin gene products show a large diversity in quantity, localization, and isoform pattern: a possible role in muscle maturation and function. J. Cell. Biol. 134, 81–92. Mas, J.A., Garcia-Zaragoza, E., Cervera, M., 2004. Two functionally identical modular enhancers in Drosophila troponin T gene establish the correct protein levels in different muscle types. Mol. Biol. Cell 15, 1931–1945. McKinsey, T.A., Zhang, C.L., Olson, E.N., 2002. Signaling chromatin to make muscle. Curr. Opin. Cell Biol. 14, 763–772. Meredith, J., Storti, R.V., 1993. Developmental regulation of the Drosophila tropomyosin II gene in different muscles is controlled by muscle-type-specific intron enhancer elements and distal and proximal promoter control elements. Dev. Biol. 159, 500–512. Nongthomba, U., Pasalodos-Sanchez, S., Clark, S., Clayton, J.D., Sparrow, J.C., 2001. Expression and function of the Drosophila ACT88F actin isoform is not restricted to the indirect flight muscles. J. Muscle Res. Cell Motil. 22, 111–119. Pennachhio, L.A., Rubin, E.M., 2001. Genomic estrategies to identify mammalian regulatory sequences. Nat. Rev. Genet. 2001;, 100–109. Reddy, K.L., Wohlwill, A., Dzitoeva, S., Lin, M.H., Holbrook, S., Storti, R.V., 2000. The Drosophila PAR domain protein 1 (Pdp1) gene encodes multiple differentially expressed mRNAs and proteins through the use of multiple enhancers and promoters. Dev. Biol. 224, 401–414. Sandelin, A., Wasserman, W.W., 2004. Constrained binding site diversity within families of transcription factors enhances pattern discovery bioinformatics. J. Mol. Biol. 338, 207–215. Scott, M.P., 1994. Intimations of a creature. Cell 79, 1121–1124. Spradling, A.C., Rubin, G.M., 1982. Transposition of cloned P elements into Drosophila germ line chromosomes. Science 218, 341–347. Stockdale, F.E., 1997. Mechanisms of formation of muscle fiber types. Cell. Struct. Funct. 22, 37–43. Sullivan, W., Ashurner, M., 2000. Drosophila Protocols. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Taylor, M.V., 1998. Muscle development: a transcriptional pathway in myogenesis. Curr. Biol. 8, R356–R358. Taylor, M.V., 2002. Muscle differentiation: how two cells become one. Curr. Biol. 12, R224–R228. Taylor, M.V., Beatty, K.E., Hunter, H.K., Baylies, M.K., 1995. Drosophila MEF2 is regulated by twist and is expressed in both the primordia and differentiated cells of the embryonic somatic, visceral and heart musculature. Mech. Dev. 50, 29–41. Vigoreaux, J.O., Saide, J.D., Valgeirsdottir, K., Pardue, M.L., 1993. Flightin, a novel myofibrillar protein of Drosophila stretch-activated muscles. J. Cell. Biol. 121, 587–598.