SEARCH

SEARCH BY CITATION

Keywords:

  • Bombyx mori;
  • chorion genes;
  • promoter architecture;
  • transcription factors

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

Regulation of silkmoth chorion genes has long been used as a model system for studying differential gene expression. The large numbers of genes, their overlapping expression patterns and the overall complexity of the system hinted towards an elaborate mechanism for transcriptional control. Recent studies, however, offer evidence of a molecular pathway governed by the interplay between two general transcription factors, CCAAT enhancer binding proteins (C/EBP) and GATA, an architectural protein, high mobility group A and a chromatin remodeller, chromo-helicase/ATPase-DNA binding protein 1. In this review we present a parsimonious model that adequately describes regulation of transcription across all temporally regulated chorion genes, and propose a role for promoter architecture.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

Eukaryotic genomes respond to environmental or physiological stimuli via multi-level control of gene expression. Such cascades provide the means for development, differentiation and adaptation (Carey et al. 2008). Control of transcriptional activation and/or repression is a critical subset of the cellular regulation machinery, deployed through an intricate circuitry of protein : protein and protein : DNA interactions. Deciphering these molecular mechanisms requires suitable systems of study. Choriogenesis of the domesticated silkmoth, Bombyx mori, is an established model system. Its key advantage is the existence of consecutive developmental phases in each single ovariole; progressively more mature follicles are found in greater distance from the germarium (Fig. 1, top). On average, each follicle is separated by ∼2 h of development from its neighbouring ones (Swevers & Iatrou, 1992).

image

Figure 1. Silkmoth ovariole and chorion gene expression patterns. Photograph of an isolated ovariole taken under the stereoscope. Follicles are positioned as in the ovary: vitellogenic closer to the germarium (left) and progressively more mature ones closer to the ooduct (right). The first choriogenic (+1) and a fully matured follicle are indicated (arrows). Choriogenic follicles are grouped (yellow boxes) according to their respective developmental stage [early (E) to very late (VL); Jones & Kafatos, 1980]. Expression patterns of different gene groups [Er.A/B, early; 6F6, early-middle (EM); L12-type, early-middle/middle; L11-type, middle/middle-late; Hc.A/B, late/very late] are depicted as RNA dot-blots (correlated with follicles in the photograph; according to Eickbush et al., 1985).

Download figure to PowerPoint

Formation of chorion, a supra-molecular protein structure, which offers mechanical protection to the oocyte, is a result of three periods of differential gene expression: early, middle and late. Chorion genes are, in accordance, characterized as early (eg Er.A/B, 6F6.1-3), middle (A/B.L12- or A/B.L11-type) or late (Hc.A/B), and their expression regulated at the level of transcriptional initiation (Fig. 1, bottom). Members of the chorion gene family reside on a continuous genetic locus on chromosome 2 (International Silkworm Genome Consortium, 2008). Contemporary α- and β-gene groups are thought to have emerged from a single ancestor gene through duplication and conversion events (Lecanidou et al., 1986). Chorion genes are, in their vast majority, organized in divergently transcribed pairs of one α- and one β-gene of the same developmental specificity, sharing a ∼250 bp-long 5′ flanking region containing all necessary cis-elements for tissue- and temporal-specific expression in vivo (Jones & Kafatos, 1980; Iatrou & Tsitilou, 1983; Spoerel et al., 1986; Spoerel et al., 1993).

Insights from dipteran oogenesis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

Dipteran chorion genes are amplified to achieve high expression levels, whereas B. mori chorion genes exist in multiple copies (Cavaliere et al., 2008). However, the very first experiments addressing chorion gene transcriptional regulation were conducted in transgenic flies. In vivo transformation of Drosophila oocytes with an early-middle silkmoth chorion gene pair promoter proved that it was sufficient for spatial and temporal control of gene expression (Mitsialis & Kafatos, 1985). Similar approaches showed that cis-elements capable of driving proper gene expression lay within a 112 bp-long region near the α-TATA box (boxed in Fig. 2A; Spoerel et al., 1993). After these experiments, it became evident that: (1) silkmoth promoters can act in both orientations, thus being bona fide bidirectional promoters, and (2) there is a common molecular context for the control of oogenesis between Diptera and Lepidoptera. More recent in vitro approaches used promoter fragments from Lepidoptera (Antheraea sp.) and Diptera (Drosophila sp., Ceratitis capitata) in gel retardation assays with silkmoth follicular nuclear extracts. These showed stage-specific protein : DNA complexes of similar (if not identical) nature to those formed by silkmoth promoters (Sourmeli et al., 2005a), a fact indicating analogies in ovarian development of species separated by 250 million years of evolution. Analogies are reflected on promoter cis-element content. For instance, in both lepidopteran and dipteran chorion gene promoters, a conserved TCACGT hexamer is identified; this is important for dipteran ovarian maturation (Fenerjian et al., 1989; Mitsialis et al., 1989; Swimmer et al., 1990; Vlachou et al., 1997). In transgenic Drosophila, deletion of the hexamer from silkmoth promoters led to loss of promoter function in either direction (Fenerjian & Kafatos, 1994); the silkmoth factor interacting with this element has yet to be identified.

image

Figure 2. Schematic representation of promoter regions. (A) Silkmoth chorion gene promoter sequence. Transcription factor binding sites are presented for the consensi of early Er.A/B, the early/early-middle 6F6, the early-middle L12-type, the middle-late L11-type and the late Hc.A/B promoter regions. Arrows above CCAAT enhancer binding protein (C/EBP) sites (oval) and GATA sites (triangle) denote orientation. Chromo-helicase/ATPase-DNA binding protein 1 (CHD1; pentagon) and high mobility group A (HMGA; white box) sites contain a question mark where a prediction for their position was made; for the 6F6 promoter no satisfactory sites were found. The TCACGT hexamer is shown as a grey box. Green boxes highlight the α-gene proximal arrays discussed in the main text. (B) The TATA-less 5′ proximal region of the C/EBP gene. Putative C/EBP, HMGA, CHD1 and GATA binding sites are shown. (C) The Drosophila melanogaster (Dm) chorion promoters of the s15 and s18 genes. Different binding sites and their orientations are presented; no data exist for HMGA, or CHD1 cis-elements. Green and red boxes highlight ‘positive’ and ‘negative’ elements for efficient gene transcription, respectively. In all panels, distances are drawn to scale (L12-type α- to β-TATA region corresponds to 211 bp); a key is also shown (left, boxed).

Download figure to PowerPoint

Other in vitro experiments using a late gene pair promoter identified two cis-elements where hypothetical transcription factors Bombyx Chorion Factor (BCF)I and II were proposed to bind (Skeiky & Iatrou, 1991). Variants of the BCFII binding site [namely a CCAAT enhancer binding protein (C/EBP) site] correspond to critical regulatory elements, characterized in vivo using the promoters of the Drosophila s15 (Mariani et al., 1988, 1996; boxed in Fig. 2C), and the silkmoth 6F6.2 chorion gene (Kravariti et al., 2001; next to the β-TATA in Fig. 2A).

The two functional faces of silkmoth C/EBP

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

Hypothetical factor BCFII recognizes a 9–10 bp-long sequence on late gene pair promoters; this, however, can be identified in all chorion promoters (with variations; Fig. 2A). The deduced consensus (TKNNGY/AAAK/C; K =T/G; Y =T/C) comprises a bipartite cis-element recognized by C/EBP belonging to the leucine zipper superfamily of transcription factors (bZIP; Miller, 2009) consisting of five families. Its members are characterized by the presence of a basic region-leucine zipper (BR-LZ) signature motif. The BR is responsible for DNA recognition and binding, the LZ mediates dimerization, a prerequisite for binding (Miller, 2009). In C/EBPs, BR-LZ motifs are found closer to the C-terminus (Vinson et al., 2002). Homo- and heterodimers are involved in various aspects of mammalian physiology (Ramji & Foka, 2002; Miller, 2009). In insects, Drosophila and B. mori homologues have been studied in respect to oogenesis (Montell et al., 1992; Rørth & Montell, 1992; Sourmeli et al., 2005a; Papantonis et al., 2008a). In the fruitfly, C/EBP control migration of slow border cells during ovarian development (Montell et al., 1992), and contribute to embryonic development (Rørth & Montell, 1992). Other C/EBP (or C/EBP-like factors) factors have been implicated in dipteran or lepidopteran physiological processes (An & Wensink, 1995; Dittmer & Raikhel, 1997; Kokoza et al., 2001; Raikhel et al., 2002; Pham et al., 2005).

B. mori chorion gene promoters carry, without exception, more than a single C/EBP recognition site. Variations of the 3′ end of the consensus allow their classification as early (TKNNGT/CAAT/G/C) or late-type (TKNNGAAAT/G/C) C/ sites, the latter having higher interaction affinity with C/EBP homodimers in vitro (Sourmeli et al., 2003; Sourmeli et al., 2005a). Results from ‘Southwestern’ analyses showed that C/EBP will bind the whole range of chorion gene promoters. In early chorion genes binding is more prominent during early (E) stages; in its interaction with early-middle or middle-late promoters the C/EBP binding peaks during early-middle (EM), middle (M) and middle-late (ML) stages, respectively (Papantonis et al., 2008a). This verifies the temporal-specific character of the C/EBP : promoter interaction. As early promoters contain (almost exclusively) early-type elements, early-middle gene pair promoters carry both early and late, and middle-late promoters solely late-type elements, can it be deduced that promoter architecture has a role in differential gene regulation?

A detailed analysis using a modified Chromatin Immunoprecipitation (ChIP) approach with staged follicles (Papantonis & Lecanidou, 2009) revealed an as yet unidentified aspect of C/EBP function. The homodimer binds cognate cis-elements on chorion promoters in two distinct phases: upon transcriptional activation of the respective gene, and after it has been turned off (Papantonis et al., 2008a). In the case of an early-middle gene pair (A/B.L9), C/EBP can be positioned on the respective promoter from EM stages (coinciding with gene activation) until M/ML stages (when transcription concludes); association with cognate sites during late choriogenesis was also observed, but did not correlate with the expression profile of the gene pair. Antisense DNA interference analysis on ex vivo developing follicles deciphered this secondary C/EBP binding event. Suppressing C/EBP expression of follicles in E stages resulted in the (foreseen) down-regulation of an early gene (Er.A1). In follicles that had reached M stages Er.A1 transcripts should not be detected. Attenuating C/EBP expression, however, led to the (unexpected) reappearance of Er.A1 transcripts (comparable to E stages; Papantonis et al., 2008a).

The interplay between chorion gene promoters and C/EBP appears to be a limiting step for initiation of gene transcription. Transcriptional termination is thought to result from the dissociation of the multi-protein complex formed. C/EBP re-binding (presumably to multiple sites) imposes a repressive state for chorion genes; changes in transcription factor availability could also play an important part. Early on during choriogenesis C/EBP is an abundant component of follicular nuclei and lower affinity protein : DNA interactions are easily favoured. During mid-choriogenesis C/EBP local concentration controls differential activation of early-middle and middle-late chorion gene pairs. The former rely upon an interaction of the C/EBP protein dimer with an early-type cis-element (lower affinity), the latter (lagging in developmental time) on a late-type:C/EBP (higher affinity) interaction (A. Papantonis and R. Lecanidou, unpubl. data; Papantonis et al., 2008a, b; Papantonis & Lecanidou, 2009). By late choriogenesis C/EBP availability is significantly lower, and we assume that the homodimer solely occupies late-type sites, thus irreversibly repressing early and middle chorion gene expression. This minimal model is supported by observations of lower affinity bZIP protein : DNA complexes being capable of protein : protein interaction deployment, whereas ‘rigid’ high affinity complexes do not favour such interplay (Lee et al., 1997; Palamarchuk et al., 2001). Nevertheless, other factors are expected to have a role in regulating gene expression.

This C/EBP-driven molecular cascade should not just rely on the ‘nature’ of cis-elements (low/high affinity) but also on the overall architecture of promoters. By examining the C/EBP promoter modules of the main chorion gene groups (Er.A/B, 6F6, A/B.L12-type, A/B.L11-type and Hc.A/B; Fig. 2A) we can deduce the following: (1) Er.A/B and 6F6 promoters contain an average of three early-type C/EBP binding sites; A/B.L12-type gene pair promoters consistently contain four sites (three early-, one late-type); A/B.L11-type gene pair promoters have four late-type sites and Hc.A/B gene pair promoters feature a single late-type element (Sourmeli et al., 2003; Fig. 1A); (2) as far as the number of C/EBP sites per bp of promoter length is concerned, the ratio is higher for early (Er.A/B, 6F6), lower for middle (A/B.L12-, A/B.L11-type) and lowest for late Hc.A/B genes; (3) aligning all promoter consensus sequences, using the β-TATA box sequence as a reference point, a GATA recognition element (AGATAA/G) holds a conserved position in (almost) all promoters, whereas the C/EBP site located between the GATA site and the β-TATA box is ‘lost’ in middle-late and late promoters. C/EBP sites near the α-TATA box also exhibit an interesting pattern: a pair of an early-type and a late-type site in early-middle (A/B.L12-type) gene promoters; a pair of late-type sites in middle-late (A/B.L11-type) ones; a single late-type C/EBP site in late (Hc.A/B) promoters in the same region as its counterpart in middle-late genes (Fig. 2A). We propose that this is indicative of an intrinsic ‘tagging’ of promoter sequences, by which they are recognized as of either earlier or later developmental specificity and accordingly regulated. C/EBP binding sites of diverse affinities correlated with diverse transcriptional outputs have been observed elsewhere (Hai & Curran, 1991; Lee et al., 1997; Palamarchuk et al., 2001). In summary, C/EBP is involved in a cross-talk with chorion promoter architecture for both activation and repression and could be the key modulator of chorion gene expression (Fig. 3).

image

Figure 3. An outline of the transcriptional circuit regulating chorion gene expression. Promoter cartoons are shown against a gradient of CCAAT enhancer binding protein (C/EBP) availability and chromatin remodelling activity (top) and a developmental time-scale [bottom; early (E) to late (L) stages]. Snapshots of regulation are shown for the C/EBP, an early (Er.A/B), a middle (L12 or L11-type) and a late (Hc.A/B) gene promoter. Chromatin remodelling occurs early during choriogenesis as a result of chromo-helicase/ATPase-DNA binding protein 1 (CHD1) activity recruited by high mobility group A (HMGA), which is found on promoters throughout choriogenesis and induces bending (only active promoters are presented in a bent conformation). It also orchestrates protein : protein interaction as follows: C/EBP antagonizes GATA for interaction with HMGA and subsequent binding to cognate cis-elements; this leads to gene activation. Genes are turned off by GATA binding and abolishing C/EBP : promoter interaction. Irreversible repression is a result of C/EBP high affinity binding to multiple sites. In this model, Hc.A/B genes do not follow the same pathway. Their up-regulation relies on a synergistic mode of action by C/EBP and GATA, recruited via HMGA. ‘T’-shaped lines denote abrogation of interaction; question marks next to transcription factor cartoons refer to lack of evidence for their function during respective stages; transcriptional status of genes (white boxes) is denoted using ‘ON’ or ‘OFF’.

Download figure to PowerPoint

GATA factors balancing between antagonism and synergy

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

The BCFI binding element from a late gene pair promoter corresponds to a GATA recognition sequence. Zinc-finger transcription factors belonging to the GATA family have been known to act as activators (Martin et al., 2001), repressors (John et al., 1996), or both (depending on relative protein levels; McNagny et al., 1998). The silkmoth BmGATAβ gene produces three isoforms, GATA1–3 (Drevet et al., 1995). The proposed model for GATA function involved its production during pre-choriogenic stages, and cytoplasmic localization in a phosphorylated form; in choriogenesis a phosphatase dephosphorylates the protein, which enters the nucleus, and triggers late chorion gene activation (Skeiky & Iatrou, 1991; Skeiky et al., 1994). The distinct expression profiles of the three BmGATAβ isoforms insinuate implication in different molecular events during oogenesis (Swevers & Iatrou, 2003).

Although GATA was initially characterized as a late gene-specific activator, its cognate binding sites are found in all chorion promoters. Gel retardation assays using early chorion gene promoters produced a complex resulting from GATA : DNA interaction, which was intensified when C/EBP-derived complexes were competed out (Sourmeli et al., 2003). Similar data of an antagonistic C/EBP-GATA relation were also produced by ChIP on middle promoters, where GATA binding correlated with down-regulation (Papantonis et al., 2008b). We hypothesize that GATA acts as a repressor early on and as an activator in later stages. In its repressive mode it antagonizes C/EBP, whereas synergy is required for late chorion gene activation.

High mobility group A, a hub for differential protein interaction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

High mobility group A (HMGA; old name: HMGI/Y) proteins comprise an abundant protein family related to chromatin and help shape its structure in vivo, hence the characterization ‘architectural nonhistone’ proteins (Bustin, 2001). HMGAs are not considered self-contained inducers of gene expression. They, however, constitute a functionally important group of factors owing to their ability to contort DNA, recognize chromatin structure with high affinity and mediate protein : protein interaction (Reeves, 2001; Foti et al., 2003; Privalov et al., 2009).

In insects, HMGA proteins have been characterized in the dipterans Drosophila and Chironomus (Wisniewski & Schulze, 1992; Claus et al., 1994), whereas lepidopteran factors have not been extensively studied (Aleporou-Marinou et al., 2003). The silkmoth HMGA (Papantonis et al., 2008b) has all characteristic features of metazoan HMGAs (small molecular mass, acidic C-terminus, AT-hook motifs), but is unique in having retained (or obtained) in the course of evolution a fourth AT-hook motif, only found in plant HMGAs (reviewed in Grasser, 2003). Another property that separates the silkmoth protein from other HMGAs is its response to phosphorylation. HMGA dynamics (ie DNA binding, eu- or heterochromatic association, interaction potential) are often regulated via reversible post-translational modifications, such as phosphorylation, methylation, or acetylation (Zhang & Wang, 2008). Casein kinase II (CK2) could phosphorylate serine residues 44/48 of HMGA in vitro, and thus disrupt protein : DNA interaction with chorion gene promoters. This actually occurs as HMGA relies on two consecutive AT-hook motifs for recognizing and interacting with A/T-rich sequences in the DNA minor groove; binding was restored upon treatment with alkaline phosphatase (Papantonis et al., 2008b). This regulation paradigm follows an ‘on/off’ mode (no intermediate response states as in mammals; Harrer et al., 2004). Finally, identification of a single HMGA transcript in follicular cells argues against multiple splice variants, another common feature of mammalian HMGAs (Reeves, 2001).

The ‘architectural’ character of HMGA during choriogenesis was highlighted by circular perturbation (or DNA bending) assays on middle gene pair promoters (A/B.L9 and A/B.L1). An average bending of ∼90° was the result of stoichiometric binding of HMGA to cognate cis-elements in vitro. Binding occurred to adjacent sequences separated by half, or one-and-a-half helix turns, and thus appears phased. Truncated HMGA versions could bind DNA, but consistently failed to reproduce the expected angle. It was therefore deduced that the first two AT-hook motifs and the acidic tail of the full-length protein were involved in efficient DNA bending (Papantonis et al., 2008b) to induce a thermodynamically preferred conformation (Privalov et al., 2009). Attempts to define the bending point on middle promoter sequences revealed that it is located between the two adjacent A/T-rich stretches that HMGA recognizes (Papantonis et al., 2008b). We propose that promoter bending contributes towards: (1) projection of particular DNA sequences to render them more accessible; (2) inducing proximity of non-adjacent cis-elements (eg HMGA : GATA interaction, discussed below).

The key role, however, of HMGA during silkmoth choriogenesis is orchestrating protein : protein interaction, which appears to dictate how activation/repression unfolds over time. According to ChIP data from middle chorion genes, the factor is bound to cognate sites throughout choriogenesis. As mentioned above, activation-related C/EBP binding sites are differentially positioned on developmentally distinct promoters. HMGA binding sites were in proximity to these, in all promoters studied (boxed in Fig. 2A). C/EBP : HMGA interaction appeared stereochemically favourable, and was verified in vivo and in vitro. C/EBP binding was directly dependent on prior HMGA binding, and C/EBP binding is of critical importance for TFIID/TBP recruitment (Papantonis et al., 2008b), signifying pre-initiation complex formation (Carey et al., 2008).

However, the HMGA : GATA interaction is evident solely during late choriogenesis, as seen by pull-down and yeast-two-hybrid assays (Papantonis et al., 2008b). GATA binds middle chorion gene promoters during ML, L choriogenic stages, its cognate sites being positioned closer to the β-TATA box (∼150 bp away from the HMGA-C/EBP cis-element cluster). Despite the physical distance separating HMGA and GATA binding sites, GATA binding in vivo is dependent upon HMGA presence on the respective promoter, and DNA bending might play a role in this (Papantonis et al., 2008b). Considering the developmental moment at which GATA is recruited to middle chorion gene promoters (when they are supposed to be turned off), and in vitro data showing an antagonistic relation between C/EBP and GATA (Sourmeli et al., 2003), we propose that the functional equilibrium between the two (controlled via interplay with HMGA) drives transcriptional activation (C/EBP-dependent) or repression (GATA-dependent) of chorion genes. This is, nonetheless, an insight from work on middle chorion genes; it still remains unclear whether this is the case for early genes as well. The HMGA : GATA interaction is also necessary for late gene (Hc.A/B) activation, in accordance to what has previously been suggested (Skeiky & Iatrou, 1991, Drevet et al., 1995). Hc.A/B genes are uniquely found in the silkmoth chorion superfamily (Kafatos et al., 1977), and an alternate pathway for their regulation would not be surprising.

Nucleosomal barriers: out of the way!

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

Many fundamental processes in eukaryotes, such as DNA replication, transcription, repair or recombination, result (or initiate) from reactions that utilize chromatin as a substrate via DNA binding proteins. A prerequisite for the identification and binding to DNA sequences is chromatin remodelling. This is executed by remodelling enzymes or complexes, and renders binding elements accessible (Clapier & Cairns, 2009). During chorion development chromo-helicase/ATPase-DNA binding protein 1 (CHD1) performs this task. It is an ATP-dependent remodeller of chromatin, and its orthologues have been isolated from various eukaryotes (Delmas et al., 1993; Stokes et al., 1996; Tran et al., 2000). They are systematically found in functional association with transcriptionally active loci (Stokes et al., 1996; Lusser et al., 2005), elongating (Simic et al., 2003; Srinivasan et al., 2005), remodelling events (Konev et al., 2007; Maier et al., 2008; Papantonis et al., 2008c) and modified histone ‘tails’ (Pray-Grant et al., 2005; Konev et al., 2007). In Drosophila, CHD1 is linked to wing organogenesis and fertility (McDaniel et al., 2008).

CHD1 expression, assessed from silkmoth EST abundance, is correlated with female insect tissues mainly related to ovarian development (eg ovary, follicular). Its chromo-domain is highly similar to its mammalian counterparts, and is predicted to recognize trimethylated lysine 4 residues on histone H3 tails (Sims et al., 2005; Marfella & Imbalzano, 2007; Flanagan et al., 2007), a signature modification of transcriptionally active chromatin (Schulze & Wallrath, 2007). These observations were experimentally confirmed using middle chorion gene pair promoters, where (1) appending of methylation marks on H3K4 and (2) C/EBP and consequent TFIID/TBP recruitment relied on CHD1 binding (Papantonis et al., 2008c; similar effects noted by Biswas et al., 2007). These events occur in strict accordance to the expression profiles of the respective genes. Moreover, CHD1 cognate cis-elements lie near the α-TATA box, adjacent to the HMGA-C/EBP cluster in middle chorion promoters (boxed in Fig. 2A), forming a long array of regulatory modules within the important promoter α-half (Spoerel et al., 1993). CHD1 binding to either A/B.L9 or A/B.L1 middle gene promoters (and Er.1A/B; A. Papantonis and R. Lecanidou, unpubl. data) occurs in vivo upon transition from vitellogenesis to choriogenesis, following (and being dependent on) HMGA recruitment (Papantonis et al., 2008b, c). In studies of the pre-choriogenic chromatin structure of the early-middle A/B.L9 gene promoter region, a nucleosome centred on the HMGA-C/EBP-CHD1 cis-element array was identified. This is an apparent physical barrier for transcription factor binding, but is translocated ∼20 bp towards the α-gene during E stages, thereby exposing the essential early-type C/EBP site. In follicles lacking CHD1 expression, C/EBP binding is obstructed as a result of nonremodelling of the respective nucleosome. The promoter half closer to the β-gene does not undergo significant remodelling of its pre-choriogenic structure (until M stages; Papantonis et al., 2008c); the GATA binding site is masked by a nucleosome, thus accessibility should be limited. This supports the presumption that GATA does not contribute to middle chorion gene activation. Data from PCRs (on staged mononucleosomal templates) covering the A/B.L1 middle-late promoter reiterated the same finding: a nucleosome on the promoter α-half is being repositioned upon transition from vitellogenesis; chromatin structure of the β-half is not affected until mid-choriogenesis (A. Papantonis & R. Lecanidou; unpubl. data). Although no evidence exists, we suspect β-proximal nucleosomes are remodelled during ML/L stages to facilitate GATA binding.

Paradigms from other organisms propose that CHD1 promotes (or hinders) transcription of only a small portion of genes per cell (2–4%; Tran et al., 2000). This seems not to be the case in follicular nuclei. After attenuating CHD1 expression (by antisense DNA interference), canonical chromatin organization is severely affected (Papantonis et al., 2008b). We estimate that a significant portion of active chromatin is functionally associated with CHD1 for this effect to be prominent. In our view, a wave of remodelling sweeps along the chorion locus, rendering critical trans-factor binding sites accessible; in turn, antagonism and/or synergy between these factors determines the transcriptional output per developmental stage (Fig. 3).

Transition from vitellogenesis to choriogenesis and hormonal signalling

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

The oogenic cascade is divided into two main phases (vitellogenesis, followed by choriogenesis), and triggered by changes in 20-hydroxylecdysone (20E) levels, a hormone produced by pupal prothoracic glands. During vitellogenesis nurse cells provide the oocyte with nutrients as yolk proteins and vitellogenin accumulate. Subsequently, the follicular epithelium builds the vitelline membrane, which surrounds the developing oocyte (reviewed in Swevers & Iatrou, 2003). Whereas 20E levels are kept low at the onset of metamorphosis (provitellogenesis), a significant elevation can be observed throughout middle and late vitellogenic stages. Interplay between 20E and its cognate receptor, Bombyx mori Ecdysone Receptor (BmEcR), is a key event of vitellogenesis (Swevers et al., 1996), but it has not been correlated with transition to choriogenic stages.

Upon transition to choriogenesis a peak in follicular cell activity is observed, as proteins that structure chorion are produced and secreted. Endogenous and exogenous factors (as regards follicles) contribute towards vitellogenic-to-choriogenic transition. Endogenous ones are also linked to the ability of vitellogenic follicles to implement their developmental programme ex vivo and produce fully matured oocytes (Swevers & Iatrou, 1992). Alongside the significant decrease in 20E levels, the presence (or absence) of different protein factors has been proposed to act as a switch for the transition: the Bombyx mori fushi tarazu factor 1 (BmFTZ-F1) nuclear receptor, or isoforms of Bombyx mori orphan nuclear receptor HR3 (BmHR3) and Bombyx mori orphan nuclear receptor E75 (BmE75) (Swevers & Iatrou, 2003). BmFTZ-F1 is expressed during late vitellogenesis (Swevers & Iatrou, 1999) and facilitates transition as in Drosophila (Zhu et al., 2006). BmFTZ-F1 activity is controlled by BmHR3, which is in turn suppressed by augmenting levels of BmE75 variants; BmE75-C and -D expression peaks as the cascade concludes. As a result, the BmHR3-BmE75 equilibrium is closely related to late vitellogenic gene regulation prior to choriogenesis initiation (Swevers et al., 2002).

The onset of choriogenesis, and thus up-regulation of chorion gene expression, results from an increase in the availability of specific transcription factors, as well as diminished expression of vitellogenesis-related genes (eg BmHR3, BmE75). Although a number of proteins are abundant components of both vitellogenic and choriogenic nuclei (ie ecdysone receptor monomers BmEcR/Bombyx mori Chorion Factor 1 (BmCF1), BmE75C, BmFTZ-F1), they have not been implicated in controlling chorion gene expression. More recent findings suggest a role for autocrine/paracrine prostaglandin signalling during follicular development; inhibiting prostaglandin biosynthesis hinders transition from vitellogenesis to choriogenesis. Under physiological conditions, prostaglandin signalling [via 3′-5′-cyclic adenosine monophosphate (cAMP)] affects follicular homeostasis, but no evidence exists linking this to the temporal-specific control of chorion genes (Machado et al., 2007).

Regulating the regulator: initial insights from the C/EBP gene

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

One of the key features at the onset of choriogenesis is the activation of the C/EBP gene. In consequence, early chorion genes are turned on via C/EBP activity, and the cascade unravels as already described (Fig. 3). The 5′ proximal region of the C/EBP gene does not contain a typical TATA-box sequence; C/EBP sites, HMGA sites and a GATA site were identified (Papantonis et al., 2008a; Fig. 2B). Data from ChIP assays show HMGA binding from the beginning of choriogenesis, and (in combination with antisense DNA interference) HMGA-dependent binding of C/EBP to its own promoter in E/EM stages; during this period C/EBP is rigorously transcribed (autoregulation via positive feed-back loops is a typical feature of bZIP-coding genes; Kockar et al., 2001; Ramji & Foka, 2002). TFIID/TBP is also recruited, following C/EBP binding, despite the absence of a TATA-box (as is the case for TATA-less promoters; eg Wright et al., 2006). GATA binding during late choriogenesis correlates with suppression of gene expression at that moment when late chorion genes are turned on (Papantonis et al., 2008b). Again, the C/EBP-GATA functional equilibrium appears to be the on/off switch of transcription.

What triggers C/EBP activation though? A plausible scenario can be based on the work of Machado et al. (2007). Prostaglandin signalling via cAMP has an effect on transition from vitellogenesis to choriogenesis (discussed above). It is established that C/EBP factors possess intrinsic cAMP-responsive activity (Wilson & Roesler, 2002) and participate in co-regulative cascades alongside cAMP response element binding (CREB) factors (cAMP response element binding protein; Nerlov, 2008; Manna et al., 2009). It would not be surprising if the silkmoth C/EBP gene reacted similarly to prostaglandin and was autoregulated by C/EBP and CREB. The silkmoth CREB orthologue produces three isoforms with expression peaks after pupation (Song et al., 2009), and a typical cAMP response element (TGACGTCA) has been identified in the C/EBP proximal promoter sequence (A. Papantonis and R. Lecanidou, unpubl. data).

Other bZIP factors and their putative roles

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

In mammals, C/EBP proteins are categorized into five groups (C/EBPα–ζ), but their orthologues are not found in all metazoans. In the silkmoth, three C/EBP-like factors have been isolated from follicular cells (in silico genome analysis did not reveal any others; Sourmeli et al., 2005a, b): C/EBP, BmCbZ and BmC/EBPγ. Their analysis, although confined to in vitro studies, did produce some results correlated with chorion gene regulation.

C/EBPγ factors are characterized by the absence of trans-activation domains; thus, their homo- or heterodimers are usually repressor molecules (Zafarana et al., 2000). Silkmoth C/EBPγ exhibits very low affinity binding for either early- or late-type C/EBP sites. It did, however, recognize a novel cis-element (gamma binding element) of an early-middle promoter. Formation of C/EBP : C/EBPγ heterodimers was not observed under physiological conditions in vitro (Sourmeli et al., 2005b). C/EBPγ did form heterodimers with CbZ (chorion bZIP protein). In fact, it is only in the context of this heterodimer that CbZ exhibits DNA binding potential. Intriguingly, the heterodimer binds typical C/EBP sites. DNase I footprinting revealed that the protein : DNA interaction covered the C/EBP recognition sequence, but also extended to adjacent A/T-rich stretches. This is because of a unique feature carried by CbZ, an AT-hook motif (as in HMGAs) precedes the BR-LZ motif and gives rise to a bipartite DNA-binding domain (Sourmeli et al., 2005b). Using this protein domain as a probe for in silico screening of available protein libraries returned a small number of homologues, exclusively from insects; CbZ appears functionally confined to insects. Given the vitellogenic expression profile of both BmCbZ and BmC/EBPγ, we speculate that the CbZ:C/EBPγ heterodimer masks C/EBP sites and renders them inaccessible up to the beginning of choriogenesis.

Future perspectives

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References

Lepidoptera, the class to which B. mori belongs, are used as model organisms in studies of molecular and developmental cell biology, molecular evolution, physiology and immunology. For B. mori in particular, availability of the complete genome sequence, of EST libraries and emerging genetic manipulation resources render the system user-friendly (Goldsmith et al., 2005; Uchino et al., 2008).

As far as choriogenesis is concerned, the presented model (outlined in Fig. 3) should not be far from the actual sequence of events governing regulation of gene expression. However, it is not nearly complete and we expect additional factors to contribute in fine-tuning gene expression. Most importantly, our hypotheses on promoter architecture and its consequent contribution to temporal specificity must undergo further investigation. We are currently developing an electroporation-based transient expression method for ex vivo developing follicles to test for that, and replace the more laborious and expensive biolistic method (Kravariti et al., 2001). The key question remains though: through what sort of a regulatory cascade are shared promoter modules sufficient for the developmentally controlled expression of a large number of genes, randomly arranged in a sole genetic locus?

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Insights from dipteran oogenesis
  5. The two functional faces of silkmoth C/EBP
  6. GATA factors balancing between antagonism and synergy
  7. High mobility group A, a hub for differential protein interaction
  8. Nucleosomal barriers: out of the way!
  9. Transition from vitellogenesis to choriogenesis and hormonal signalling
  10. Regulating the regulator: initial insights from the C/EBP gene
  11. Other bZIP factors and their putative roles
  12. Future perspectives
  13. Acknowledgements
  14. References