InteractiveFly: GeneBrief

vielfaltig: Biological Overview | References

In all animals, the initial events of embryogenesis are controlled by maternal gene products that are deposited into the developing oocyte. At some point after fertilization, control of embryogenesis is transferred to the zygotic genome in a process called the maternal-to-zygotic transition. During this time, many maternal RNAs are degraded and transcription of zygotic RNAs ensues. There is a long-standing question as to which factors regulate these events. The recent findings that microRNAs and Smaug mediate maternal transcript degradation have shed new light on this aspect of the problem. However, the transcription factor(s) that activate the zygotic genome remain elusive. The discovery that many of the early transcribed genes in Drosophila share a cis-regulatory heptamer motif, CAGGTAG and related sequences, collectively referred to as TAGteam sites raised the possibility that a dedicated transcription factor could interact with these sites to activate transcription. This study reports that the zinc-finger protein Zelda (Zld; Zinc-finger early Drosophila activator) binds specifically to these sites and is capable of activating transcription in transient transfection assays. Mutant embryos lacking zld are defective in cellular blastoderm formation, and fail to activate many genes essential for cellularization, sex determination and pattern formation. Global expression profiling confirmed that Zld has an important role in the activation of the early zygotic genome and suggests that Zld may also regulate maternal RNA degradation during the maternal-to-zygotic transition (Liang, 2008).

In Drosophila, an initial wave of zygotic gene transcription occurs between 1 and 2h of development during the mitotic cleavage cycles 8-13. This is followed by a major burst of activity between 2 to 3h of development (cycle 14) when the embryo is undergoing cellular blastoderm formation. Many pre-cellular genes contain TAGteam sites in their upstream regulatory regions including several direct targets of Bicoid, Dorsal and other key regulators of patterning (ten Bosch, 2006; De Renzis, 2007; Li, 2008). It has been previously demonstrated (ten Bosch, 2006) that TAGteam sites are required for the early expression of the dorsoventral gene zen, and the sex determination genes sisB (also known as sc) and Sxl. To isolate the TAGteam binding factor, a yeast one-hybrid screen was performed with a 91 base-pair (bp) fragment from the zen cis-regulatory region (zen(91)), which contains four TAGteam sites. zld, encoded by the X chromosomal gene CG12701 (also known as vfl), was selected as the only candidate of the 11 recovered that had the potential to bind specific DNA sequences because it encoded a protein with six C2H2 zinc fingers. Oligonucleotides with different TAGteam sites were tested in gel shift assays with the 357 amino acid carboxy-terminal region of Zld fused to glutathione S-transferase (GST-ZldC). All oligonucleotides tested formed complexes with GST-ZldC, although with different affinities, whereas mutations in the heptanucleotide sequence abolished binding. Notably, the site with the strongest affinity, CAGGTAG, is the site most over-represented in regulatory elements of pre-blastoderm genes versus post-blastoderm genes (ten Bosch, 2006). A plasmid expressing full-length Zld protein promoted transcriptional activation of a zen(91)-lacZ reporter but not a mutated zen(91m)-lacZ reporter after co-transfection in Drosophila S2 cells. Taken together, these data strongly suggest that Zld activates transcription of zen and probably other TAGteam-containing genes (Liang, 2008).

zld transcripts were detected in the germline cells of the ovary, in unfertilized eggs, and throughout early development. Later, zld becomes restricted to the nervous system and specific head regions, as previously shown (Staudt, 2006). To analyse zld function, deletion alleles of zld were generated by imprecise excision. Hemizygous embryos showed abnormal central nervous system and head development, consistent with previous reports of CG12701 lethal P-insertion phenotypes (Staudt, 2006). zld transcripts were not observed in these embryos after cycle 14. However, younger embryos had high levels of maternal zld transcripts, indicating that maternally loaded zld transcripts are degraded during cellularization and replaced with zygotic zld (Liang, 2008).

To eliminate maternal zld from embryos, clones of zld²⁹⁴ mutant germ cells were induced in the adult female. All resulting embryos were null for maternal zld (M^- zld), and the male embryos were also null for zygotic zld (M^-Z^- zld). All early M^- zld embryos lacked zld transcripts but had normal patterns of other maternally deposited factors such as bicoid RNAs and the Dorsal protein gradient. Unlike M^-Z⁺ zld embryos, which began to express zld ubiquitously in cycle 14, M^-Z^- zld embryos never expressed zld. However, regardless of their zygotic genotypes, all M^- zld embryos showed a severely abnormal morphology after cycle 14 and did not survive to make cuticle (Liang, 2008).

Before cycle 14, M^- zld embryos are similar to wild type, except for sporadic nuclear fallout. However, at early cycle 14 the hexagonal-actin network becomes disorganized and begins to degenerate resulting in a multinucleated phenotype resembling nullo and Serendipity α (Sry-α) mutants. Cellularization does not proceed as furrow canals never move inward like in the wild type, and Neurotactin (Nrt) accumulates abnormally in the apical cytoplasm -- reminiscent of the slow as molasses (slam) mutant phenotype. Staining with anti-Slam antibody confirmed that Slam protein is mostly absent by mid-cycle 14, whereas Slam has moved basally in wild type. In addition, nuclei do not elongate but instead become rounded, enlarged and clump together. Regions of higher nuclear density were observed, a phenotype similar to that obtained by injection of CG12701 double-stranded RNA (Staudt, 2006), which resembles a frühstart (frs, also known as Z600) phenotype (Grosshans, 2003). Despite their aberrant morphology, M^- zld embryos attempt to form a ventral furrow but soon become highly disorganized with only pole cells recognizable. The M^- zld cellularization defects were rescued by driving a wild-type copy of zld into the germ line using the ovarian tumour (otu) promoter. The cytoskeleton becomes well structured and furrow canal ingression is normal as Slam protein is restored (Liang, 2008).

The broad range of phenotypes strongly indicated that M^- zld embryos do not express genes essential for cellular blastoderm formation. Expression was examined of Sry-α, slam and nullo, as well as sisA, sisB, sisC (also known as os), Sxl, zen and dpp. None of these genes was activated in M^- zld embryos, except at the poles in some cases. However, sna and sog (which are activated by Dorsal) were not absent but were delayed in expression by at least two cycles, suggesting that Zld facilitates the onset of early gene transcription. Furthermore, the lateral stripes of sog were greatly reduced in width, indicating that in regions in which Dorsal protein amounts are low, a combinatorial mechanism involving both Dorsal and Zld establishes the broad sog domain. Notably, there are two TAGteam sites in the 393 bp sog enhancer that lie close to Dorsal binding sites (Liang, 2008).

The results indicated that Zld is a global activator of early genes. To test this directly the expression profiles of wild-type and M^- zld embryos were compared in mitotic cycles 8-13, a time point presumably enriched in genes that are direct Zld targets. One-hundred-and-twenty genes were downregulated and surprisingly 176 genes were upregulated at least twofold, in the absence of Zld. The downregulated set was strongly enriched in genes that are zygotically expressed and involved in early developmental processes, including most of the genes assayed in situ. For example, 12 genes involved in cellular blastoderm formation (nullo, slam, Sry-α, bnk, frs, btsz, halo and 5 halo-like genes (Gross, 2003), 6 sex-determination genes (sisA, sisB, sisC, run, Sxl and dpn), and 8 dorsoventral genes (dpp, tld, tok, tsg, tsg-like, scw, zen and zen2) are in the downregulated data set. Overall, 75% of the early genes previously described as pre-cellular are included. This number may be an underestimate because there could be many genes such as sna and sog that did not make the twofold cutoff but are indeed regulated by Zld (Liang, 2008).

About 80% of the downregulated genes have TAGteam sites within 2 kilobases (kb) upstream of the transcription start site, and another 10% have TAGteam sites in introns, such as slam with two sites in its first intron, supporting the idea that most of the downregulated genes are direct Zld targets. In addition, the TAGteam sites upstream of the downregulated genes tend to be located very close to the transcription start site within 200 bp, consistent with the previous finding that early zygotic genes have a statistical over-representation of TAGteam sites close to the start site (Liang, 2008).

In contrast to the downregulated genes, the upregulated set is strongly enriched in genes that are maternally expressed. The possibility was considered that Zld activates components of the RNA degradation machinery that in turn destabilize maternal RNAs. Because the microRNA (miRNA) miR-309 enhancer contains two TAGteam sites, miR-309 primary transcripts were assayed in M^- zld embryos, and indeed they were absent. It was recently shown that mature miR-309 miRNAs become abundant during cycle 14, and are involved in maternal transcript turnover in 2-4 h embryos. Not surprisingly, the 1-2 h (cycles 8-13) data set had no overlap with the 44 published miR-309 targets, however 2-4 h profiling experiments should demonstrate whether they are upregulated in the absence of zld. The upregulated genes were compared to those affected by smaug, another gene required for the removal of maternally supplied RNAs. There was little overlap with the published Smaug targets, suggesting that Zld is involved in a parallel pathway of maternal RNA degradation (Liang, 2008).

This study has demonstrated that Zld functions as a key transcriptional activator during the maternal-to-zygotic transition (MZT) in Drosophila. This is the first demonstration of such an activator in any organism. It is proposed that the biological role of Zld in the pre-blastoderm embryo is to set the stage for vital processes such as cellular blastoderm formation, counting of X chromosomes for dosage compensation and sex determination, and pattern formation, by ensuring the coordinated accumulation of batteries of gene products during the MZT. This early preparedness should allow sufficient time for the formation of molecular machines involved in these processes, and so are ready to spring into action during the prolonged interphase of cycle 14 (Liang, 2008).

Mutations of the Drosophila zinc finger-encoding gene vielfältig impair mitotic cell divisions and cause improper chromosome segregation.

This study describes the molecular characterization and function of vielfältig (vfl), a X-chromosomal gene that encodes a nuclear protein with six Krüppel-like C2H2 zinc finger motifs. vfl transcripts are maternally contributed and ubiquitously distributed in eggs and preblastoderm embryos, excluding the germline precursor cells. Zygotically, vfl is expressed strongly in the developing nervous system, the brain, and in other mitotically active tissues. Vfl protein shows dynamic subcellular patterns during the cell cycle. In interphase nuclei, Vfl is associated with chromatin, whereas during mitosis, Vfl separates from chromatin and becomes distributed in a granular pattern in the nucleoplasm. Functional gain-of-function and lack-of-function studies show that vfl activity is necessary for normal mitotic cell divisions. Loss of vfl activity disrupts the pattern of mitotic waves in preblastoderm embryos, elicits asynchronous DNA replication, and causes improper chromosome segregation during mitosis (Staudt, 2006).

Evidence is provided that the C2H2 zinc finger protein Vfl participates in mitotic cell division and eventually causes abnormal chromosome segregation during mitosis. This defect results in distinct organismal phenotypes, which are most prominently demonstrated by the lack of synchronous mitotic waves during early Drosophila embryogenesis, as reflected by uncoordinated division patterns and asynchronous DNA-replication cycles. As a result, preblastoderm nuclei are no longer arranged in a single layer at the periphery of the embryo and cause the formation of a number of extra folds during gastrulation. Subsequently, embryos develop an aberrant segmentation pattern, a defective nervous system, and an abnormal muscle pattern in the developing embryo. The mitosis-related phenotype of the vfl mutants is consistent with the notion that vfl transcripts are highly enriched in mitotically active cells and that the protein exerts a dynamic subcellular localization pattern during the cell cycle. From early telophase until the end of interphase, the Vfl protein is chromatin associated. During the DNA-condensation phase, it dissociates from chromatin, remains separated from DNA during metaphase and anaphase, and accumulates in a granular pattern in the area of the disintegrated nucleus once its envelope is dissolved. At this stage, Vfl is neither associated with lamin nor with microtubules or DNA. At the end of mitosis and upon entering the telophase, the granules vanish, and Vfl becomes instantly associated with interphase chromatin. This observation and the canonical DNA-binding domain of Vfl suggest that the protein is active when bound to DNA and inactive when dissociated from chromatin. This proposal implies that the low amounts of Vfl in the cytoplasm of mitotically inactive cells are likely to represent a storage of inactive protein and that Vfl functions once it is chromatin associated during interphase of dividing cells. Vfl may act as a DNA-binding factor that participates directly in chromatin dynamics and accessibility or act indirectly as a transcription factor that regulates the expression of genes involved in these processes. However, because the loss of vfl activity causes notable defects before mitosis 14, i.e., the time point when the embryo switches from maternally controlled nuclear divisions to the zygotic control of mitosis, it seems unlikely that Vfl acts as a transcription factor. Thus, the idea is favored that Vfl functions in DNA replication or participates in some aspects of chromatin dynamics required for proper mitosis (Staudt, 2006).

BrdU labeling experiments of vfl mutant and wild-type embryos indicate that the coordinated timing of mitosis, including the process of DNA replication, is strongly impaired by altering the normal dose of vfl activity. However, DNA replication is not blocked in the mutants as indicated by the appearance of up to four nuclei that remain associated by chromatin bridges, indicating that the partially separated nuclei are capable to undergo replication although sister chromatids have failed to separate during the previous anaphase. These observations and the early association of Vfl with early interphase chromatin suggest that the protein could participate in the temporal control of DNA replication. Alternatively, or in addition, the early chromatin association of Vfl could also reflect its requirement during DNA decondensation and/or cohesin loading, two processes that are prerequisite for the initiation of replication (Staudt, 2006).

An interesting aspect of the results concerns the regulation of Vfl localization during mitosis, i.e., association of Vfl with interphase chromatin and with granular structures during all other phases of mitosis. Chromatin condensation during prophase, when Vfl dissociates from chromatin, is accompanied by the cessation of transcription. This process correlates with inactivation of transcriptional regulators as well as other regulatory chromatin components by removing them from their DNA targets. Such an in vivo mechanism has been recently reported for a mammalian zinc finger protein Ikaros, which plays a key role in the development and the response of the immune system. Ikaros has also been implicated in the regulation of cell cycle progression and was found to dissociate from chromatin during early stages of mitosis due to a G₂/M-specific phosphorylation event. It is not known how the dissociation of Vfl from chromatin and its accumulation in the granular structures are achieved mechanistically. It is speculated, however, that these structures represent the mitotic containment or a sequestering form for Vfl until cells enter the interphase again. In contrast, cells that do not continue mitosis nevertheless maintain comparatively low levels of Vfl in cytoplasmic granules, likely to represent a small pool of stored and inactive Vfl protein (Staudt, 2006).

Vfl has no direct vertebrate homologue that could be identified by sequence comparison. However, this result has to be taken with caution due to some specifics of zinc finger domains. The C2H2 zinc finger motif is an unusually small, self-folding domain of 25- to 30-amino acid residues. It includes paired cysteines and histidines as zinc-coordinating residues and possesses two short β-strands followed by an α-helix. The DNA-binding properties usually depend on no more than three amino acid residues of the zinc finger loop, the arrangement of C2H2 domains in proteins, and the higher order structure of proteins. This arrangement and that only a few conserved amino acid residues are required to ensure the sequence-specific DNA binding make it extremely difficult, if not impossible, to predict how the homologous vertebrate C2H2 finger would look. In contrast, the protein was found to be highly conserved in all insects analyzed, including the mosquito An. gambiae and the beetle T. castaneum, which separated from Drosophila for more than 250 and 300 million years ago, respectively. Thus, the function of Vfl might be still unrecognized among the more than 2000 C2H2 zinc finger proteins that were annotated in the mouse and human genomes. Alternatively, vfl may have a function for the insect specific nuclear divisions during syncytial blastoderm stage and therefore may not be conserved in species other than insects (Staudt, 2006).

Loss of vfl activity causes abnormally shaped and enlarged nuclei that fail to become integrated into the cortical arrangement of preblastoderm nuclei. Obviously, these morphological features, in particular the enlargement of the nuclei, cause some space constrictions at the periphery of the embryo, and thus a significant portion of nuclei fail to align properly. The subsequent divisions, which occur in part perpendicular to the normal division axis as frequently observed in such embryos, cause an additional space limitation which forces the epithelial cell layer to form the irregular and variable patterns of folds that were consistently observed in gastrulating embryos. Notably, the positions of the folds vary significantly from embryo to embryo. This finding can be attributed to the fact that the embryos analyzed were most likely not lack-of-function ('null') mutants and always contained half the normal dose of maternal vfl activity. Furthermore, although the RNAi injections into embryos were done as early as possible after egg deposition, some undetected amounts of protein might have already accumulated in such embryos. This proposal is consistent with the notion of earlier and more severe effects in response to increasing amounts of injected RNAi. Irrespectively of this speculation, the results presented in this study show that inadequate levels of Vfl interfere with the timing of mitosis and eventually result in impaired chromosome separation. The specific expression of vfl in mitotically active cells, the dose dependence of the protein as demonstrated by gain-of-function and loss-of-function experiments and the shuttling of Vfl between chromatin of interphase nuclei and the granular structures during the other stages of the mitotic cycle suggest that protein function is tightly regulated. The mechanisms of the shuttling, the molecular pathways in which Vfl participates, and the link between Vfl activity and DNA replication remain to be further elucidated (Staudt, 2006).

Design flexibility in cis-regulatory control of gene expression: synthetic and comparative evidence

In early Drosophila embryos, the transcription factor Dorsal regulates patterns of gene expression and cell fate specification along the dorsal-ventral axis. How gene expression is produced within the broad lateral domain of the presumptive neurogenic ectoderm is not understood. To investigate transcriptional control during neurogenic ectoderm specification, divergence and function of an embryonic cis-regulatory element controlling the gene short gastrulation (sog) was studied. While transcription factor binding sites are not completely conserved, it has been demonstrated that these sequences are bona fide regulatory elements, despite variable regulatory architecture. Mutation of conserved sequences revealed that putative transcription factor binding sites for Dorsal and Zelda, a ubiquitous maternal transcription factor, are required for proper sog expression. When Zelda and Dorsal sites are paired in a synthetic regulatory element, broad lateral expression results. However, synthetic regulatory elements that contain Dorsal and an additional activator also drive expression throughout the neurogenic ectoderm. These results suggest that interaction between Dorsal and Zelda drives expression within the presumptive neurogenic ectoderm, but they also demonstrate that regulatory architecture directing expression in this domain is flexible. A model for neurogenic ectoderm specification is proposed in which gene regulation occurs at the intersection of temporal and spatial transcription factor inputs (Liberman, 2009).

Through a comparative analysis of orthologous sog cis-regulatory modules from twelve Drosophilid species, core regulatory elements conserved in these sequences were identified. Considerable binding site turnover has occurred during the approximately 40 million years of evolution, yet some sequences are conserved. This observation supported the hypotheses that were investigated in this work which are, 1) that conserved sequences are functionally required and, 2) that variable architectures might generate the same or similar patterns of expression. Surprisingly, despite the opportunity for binding site turnover during the course of evolution, the sog regulatory regions from D. virilis can still be interpreted faithfully when used to drive reporter expression in D. melanogaster. It is concluded from these experiments, despite flexibility in the cis-regulatory element structure, regulatory logic has been conserved during evolution of the cis-regulatory module sequences to support sog expression (Liberman, 2009).

Though this comparative analysis identified limited sequence homology, what sequence conservation that was present facilitated efforts to examine the core regulatory elements required for patterning the neurogenic ectoderm. Using site-directed mutagenesis to eliminate sites within the sog cis-regulatory sequence, results were obtained that suggest that Dorsal functions together with the ubiquitous activator Zelda to control sog expression within the neurogenic ectoderm. Furthermore, synthetic cis-regulatory elements were constructed, consisting of Dorsal and Zelda or Dorsal and D-STAT sites, which are both able to support expression in the broad lateral domain of Drosophila early embryo. From these results it is concluded that broad lateral expression is achieved by a combination of Dorsal sites and sites for the ubiquitous activator Zelda, which suggests that a more general mechanism to create broad expression may involve interactions between Dorsal and other broadly expressed transcription factors (Liberman, 2009).

Mutagenesis and mutant analysis results demonstrate that Dorsal and Zelda support expression of sog along the dorsal-ventral axis. In the absence of Dorsal protein, expression of sog is gone; however when Dorsal binding sites were mutagenized, weak ventral-lateral reporter expression remains that could be due to unknown Dorsal binding sites that were not detected by PWM searches or due to input from another transcription factor. In the absence of Zelda binding sites or in Zelda mutants, expression is slightly broader than when Dorsal sites are eliminated. This residual expression could be due to Dorsal and/or another transcription factor (e.g. bHLH) functioning to direct expression, in a Zelda-independent manner, within the ventral-neurogenic ectoderm; however, the data suggests that Twist is not likely involved, as the domain of sog expression along the dorsal-ventral axis is not severely affected in twist mutants (Liberman, 2009).

Previous genetic studies have demonstrated that Dorsal is required for specification of the presumptive neurogenic ectoderm, but binding sites for Dorsal alone are not sufficient to generate expression within the broad lateral domain of embryos. Dorsal has been shown to function synergistically with Twist to pattern the presumptive mesoderm and ventral neurogenic ectoderm. This study presents evidence that Dorsal and Zelda function synergistically to regulate expression that is able to encompass the entire presumptive neurogenic ectoderm domain. Some method of cooperativity likely exists between Dorsal and Zelda, at the level of DNA binding or downstream, and is responsible for extending the expression domain into dorsal-lateral regions of the embryos, where the levels of nuclear Dorsal are low (Liberman, 2009).

It is proposed that Dorsal functions as a spatial regulator in the neurogenic ectoderm and that additional transcription factors like Zelda, act as co-activators to regulate the precise onset of expression. Furthermore, it is suggested that multiple ubiquitous or broadly expressed activators may function with Dorsal to support expression in a broad lateral domain (e.g. Zelda, STAT, and bHLH transcription factors such as Daughterless (Da). This study has demonstrated that STAT binding sites can also function together with Dorsal to drive expression in a broad lateral domain. Further support for this idea includes the observation that sog as well as ths exhibit broad expression early. Sites for Zelda are also present in the ths cis-regulatory module, and these sites likely direct the almost-ubiquitous early expression of ths observed. Interaction of Dorsal with distinct co-activators may not only regulate the spatial domain of expression supported, but also the temporal output. Zelda along with Dorsal or a Dorsal target initiates the earliest zygotic expression detected; perhaps interactions between Dorsal and other activators facilitate expression within a broad lateral domain (or other defined pattern) at later time-points. It is asserted that gene expression is achieved at the intersection of the Dorsal nuclear gradient and the additional activator which could either be ubiquitous in the case of Zelda or localized in the case of Twist (Liberman, 2009).

Even equipped with this new knowledge, other cis-regulatory modules that support co-expression of genes SoxN, pyramus and Neu3 have proven difficult to identify. To date, SoxN and pyramus regulatory elements remain unidentified. Flexible regulatory structures could account for some of the obscurity that has been encountered in the identification of cis-regulatory modules that support expression of genes within Drosophila early embryos. Flexibility in binding site composition, orientation and number of sites has also been demonstrated in the regulation of co-expressed genes in Ciona by extensive co-expression analyses. Possibly the observed flux in binding site composition and arrangement provides a mechanism that facilitates the introduction of mutations, which may be selected when a fitness advantage is provided to the developing embryo (Liberman, 2009).

Recently, a second regulatory element for sog located upstream of the gene was identified which also drives expression in a broad lateral stripe in the presumptive neurogenic ectoderm of cellularized embryos. This novel regulatory element as well as the known regulatory element, the intronic enhancer examined in this study, probably function together to control the full expression pattern of sog in the developing embryo. While both cis-regulatory sequences contain Dorsal and Zelda binding sites, the novel enhancer contains many more bHLH sites (L. Liberman, unpub. obs.), which is in stark contrast to the intronic sog regulatory element, which contains only one bHLH site and exhibits very little change of expression in twist mutant embryos. This new regulatory element presents further evidence that there exist multiple solutions for the developmental problem of producing spatially and temporally regulated expression. Future experiments will address whether these early embryonic enhancers controlling the expression of the sog gene within similar domains use the same mechanism (i.e. Dorsal + Zelda cooperativity) to support expression in a broad lateral stripe or whether different mechanisms are used (Liberman, 2009).

Evolutionary comparisons of sequences from diverged species can be very useful for the dissection of underlying cis-regulatory logic, as has been shown in this study; yet the important variable is that the proper comparisons of sequences must be made (i.e. species of appropriate evolutionary distance) and this is not always easy to define. In vertebrate systems, analyses of cis-regulatory modules usually focus on modules identified by methods that select for high degrees of conservation, which inherently have a low amount of flexibility. Arguments have been made that deciphering the underlying regulatory logic from evolutionary comparisons of sequences, when conservation is too high, is hard to interpret. However, it is contended that the relevant comparisons that provide insights into cis-regulatory logic are context-dependent. In analysis of the sog and Neu3 cis-regulatory modules, only limited sequence conservation was identified in comparisons of homologous sequences isolated from D. melanogaster and other Drosophilids. In the sog early embryonic regulatory element that was analyzed in this study, 71 (of 395) base-pairs of non-contiguous sequence exhibits conservation. The degree of conservation that was retained however was useful for dissecting the underlying regulatory logic (Liberman, 2009).

Identifying regulatory regions with flexible structure is more challenging than scanning for a stringent set of binding sites, but it may also reveal alternative mechanisms for specification that were not previously considered. It is predicted that studies that dissect the flexibility of cis-regulatory modules may one day provide insights to facilitate dissection of vertebrate regulatory elements in general, including ones that exhibit flexibility of sequence. It seems plausible that stringently conserved regulatory elements control gene expression of certain classes of genes, like those required for certain essential processes. Flexible regulatory architectures may provide a mechanism for generating variability throughout evolution. Ultimately it will prove useful to make evolutionary comparisons with both highly conserved sites and flexible architectures to determine how each contributes to establishment or maintenance of gene regulation (Liberman, 2009).

Combinatorial activation and concentration-dependent repression of the Drosophila even skipped stripe 3+7 enhancer

Despite years of study, the precise mechanisms that control position-specific gene expression during development are not understood. This study analyzed an enhancer element from the even skipped (eve) gene, which activates and positions two stripes of expression (stripes 3 and 7) in blastoderm stage Drosophila embryos. Previous genetic studies showed that the JAK-STAT pathway is required for full activation of the enhancer, whereas the gap genes hunchback (hb) and knirps (kni) are required for placement of the boundaries of both stripes. The maternal zinc-finger protein Zelda (Zld) is absolutely required for activation, and evidence is presented that Zld binds to multiple non-canonical sites. A combination of in vitro binding experiments and bioinformatics analysis was used to redefine the Kni-binding motif, and mutational analysis and in vivo tests to show that Kni and Hb are dedicated repressors that function by direct DNA binding. These experiments significantly extend understanding of how the eve enhancer integrates positive and negative transcriptional activities to generate sharp boundaries in the early embryo (Struffi, 2011).

The experiments described in this study significantly refine understanding of how the eve 3+7 enhancer functions in the early embryo. In particular, it was shown that the maternal zinc-finger protein Zld is absolutely required for STAT-mediated enhancer activation, and that the gap proteins Kni and Hb establish stripe boundaries by directly binding to multiple sites within the enhancer (Struffi, 2011).

When first activated in late nuclear cycle 13, the minimal eve 3+7 enhancer drives weak stochastic expression in a broad central pattern, which refines in cycle 14 to a stripe that is about four nuclei wide. By contrast, stripe 7 expression, which is visible by enzymatic staining methods, is nearly undetectable using fluorescence in situ hybridization (Struffi, 2011).

Previous work showed that stripe 7 shares regulatory information with stripe 3 but is also controlled by sequences located between the minimal stripe 3+7 and stripe 2 enhancers, and possibly by sequences within and downstream of the stripe 2 enhancer. Thus, stripe 7 is unique among the eve stripes in that it is not regulated by a discrete modular element (Struffi, 2011).

Previous work showed that the terminal gap gene tailless (tll) is required for activation of eve 7. However, since the Tll protein probably functions as a dedicated repressor, it is likely that activation of eve 7 by Tll occurs indirectly, through repression of one or more repressors (Struffi, 2011).

The ubiquitous maternal protein Zld is required for the in vivo function of both the eve 3+7 and eve 2 enhancers, which are activated by the JAK-STAT pathway and Bicoid (Bcd), respectively. Zld was previously shown to bind to five sequence motifs that are over-represented in the regulatory regions of early developmental genes. Mutations of the single TAGteam site in the eve 3+7 enhancer caused a reduction in expression, but zld M- embryos, mutant for maternal zld expression, showed complete abolishment of eve 3+7-lacZ reporter gene expression. Also, the eve 2 enhancer, which does not contain any canonical TAGteam sites, is nonetheless inactive in zld M- embryos. This study showed that this enhancer contains at least four variants of the TAGteam sites, which suggests that Zld binding to non-canonical sites is crucial for its function in embryogenesis. ChIP-Chip data show that Zld binding extends throughout much of the eve 5' and 3' regulatory regions (Struffi, 2011).

The implication of such broad binding and the requirement for Zld for activation of two eve enhancers are consistent with its proposed role as a global activator of zygotic transcription. How might this work? One possibility is that there are cooperative interactions between Zld and the other activators of these stripes. A non-exclusive alternative is that Zld binding creates a permissive environment in broad regions of the genome, possibly by changing the chromatin configuration and making it more likely that the other activator proteins can bind. However, it is important to note that eve expression is not completely abolished in zld M- embryos, so at least some eve regulatory elements could function in the absence of Zld. Future experiments will be required to further characterize the role of Zld in the regulation of the entire eve locus (Struffi, 2011).

The genetic removal of kni causes a broad expansion of eve 3+7- lacZ expression in posterior regions of the embryo, and ectopic Kni causes a strong repression of both stripes. Interestingly, the posterior boundary of eve stripe 3 is positioned in regions with extremely low levels of Kni protein. If the stripe 3 posterior boundary is solely formed by Kni, the enhancer must be exquisitely sensitive to its repression, possibly through the high number of sites in the eve 3+7 enhancer. Previous attempts to mutate sites based on computational predictions failed to mimic the genetic loss of kni, so this study used a biochemical approach to identify Kni sites in an unbiased manner. EMSA analyses identified 11 Kni sites, and the PWM derived from these sites alone is very similar to the Kni matrix derived in a bacterial one-hybrid study. Thus, these studies provide biochemical support for the bacterial one-hybrid method as an accurate predictor of the DNA-binding activity of this particular protein (Struffi, 2011).

It was further shown that specific point mutations abolish binding to nine of the 11 sites, and when these mutations were tested in a reporter gene they caused an expansion that is indistinguishable from that detected in kni mutants. This result strongly suggests that Kni-mediated repression involves direct binding to the eve 3+7 enhancer, and that Kni alone can account for all repressive activity in nuclei that lie in the region between stripes 3 and 7. However, this work does not address the exact mechanism of Kni-mediated repression. The simplest possibility is that Kni competes with activator proteins for binding to overlapping or adjacent sites. This mechanism is considered unlikely because only one of the 11 Kni sites overlaps with an activator site. Also, the in vivo misexpression of a truncated Kni protein (Kni 1-105) that contains only the DNA-binding domain and the nuclear localization signal has no discernible effect on the endogenous eve expression pattern, whereas a similar misexpression of Kni 1-330 or Kni 1-429 strongly represses eve 3+7 (Struffi et al., 2004) (Struffi, 2011).

Whereas Kni-mediated repression forms the inside boundaries of the eve 3+7 pattern, forming the outside boundaries is dependent on Hb, which abuts the anterior boundary of stripe 3 and overlaps with stripe 7. Both stripes expand towards the poles of the embryo in zygotic hb mutants, and these expansions are mimicked by mutations in four or all nine Hb sites within the eve 3+7 enhancer. Further anterior expansions of the pattern are prevented by an unknown Bcd-dependent repressor (X) and the Torso (Tor)-dependent terminal system. Indeed, eve 3+7-lacZ expression expands all the way to the anterior tip in mutants that remove bcd and the terminal system (Struffi, 2011).

The mutational analyses suggest that Hb is a dedicated repressor of the eve 3+7 enhancer, and argue against a dual role in which high Hb levels repress, whereas lower concentrations activate, transcription. One caveat is that activation of the stripe might occur via maternal Hb in the absence of zygotic expression. However, triple mutants that remove zygotic hb, kni and tor, a terminal system component, show eve 3+7 enhancer expression that extends from ~75% embryo length (100% is the anterior pole) to the posterior pole. It is extremely unlikely that the maternal Hb gradient, which is not perturbed in this mutant combination, could activate expression throughout the posterior region. It is proposed that any activating role for Hb on this enhancer is indirect and might occur by repressing kni, which helps to define a space where the concentrations of both repressors are sufficiently low for activation to occur. kni expands anteriorly in hb mutants and is very sensitive to repression by ectopic Hb, consistent with an indirect role in activation. A similar mechanism has been shown to be important for the correct positioning of eve stripe 2. In this case, the anterior Giant (Gt) domain appears to be required for eve 2 activation, but it does so by strongly repressing Kr, thus creating space for activation in the region between Gt and Kr (Struffi, 2011).

The correct ordering of gene expression boundaries along the AP axis is crucial for establishing the Drosophila body plan. All gap genes analyzed so far seem to function as repressors that differentially position multiple boundaries. However, it is still unclear how differential sensitivity is achieved at the molecular level. Simple correlations of binding site number and affinity with boundary positioning cannot explain the exquisite differences in the sensitivity of individual enhancers, suggesting that they do more than 'count' binding sites and that specific arrangements of repressor and activator sites might control this process. The experiments described here better define the binding characteristics of both Hb and Kni and provide a firm foundation for future experiments designed to decipher the regulatory logic that controls differential sensitivity (Struffi, 2011).

De Renzis, S. D., Elemento, O., Tavazoie, S. and Wieschaus, E. F. (2007). Unmasking activation of the zygotic genome using chromosomal deletions in the Drosophila embryo. PLoS Biol. 5: 1036-1051. PubMed Citation: 17456005

Gross, S. P., Guo, Y., Martinez, J. E. and Welte, M. A. (2003). A determinant for directionality of organelle transport in Drosophila embryos. Curr. Biol. 13: 1660-1668. PubMed Citation: 14521831

Grosshans, J., Müller, H. and Wieschaus, E. (2003). Control of cleavage cycles in Drosophila embryos by frühstart. Dev. Cell 5: 285-294. PubMed Citation: 12919679

Li, X. et al. (2008). Transcription factors bind thousands of active and inactive regions in the Drosophila blastoderm. PLoS Biol. 6: 365-388. PubMed Citation: 18271625

Liang, H. L., Nien, C. Y., Liu, H. Y., Metzstein, M. M., Kirov, N. and Rushlow, C. (2008). The zinc-finger protein Zelda is a key activator of the early zygotic genome in Drosophila. Nature 456(7220): 400-3. PubMed Citation: 18931655

Liberman, L. M. and Stathopoulos, A. (2009). Design flexibility in cis-regulatory control of gene expression: synthetic and comparative evidence. Dev. Biol. 327(2): 578-89. PubMed Citation: 19135437

Staudt, N., Fellert, S., Chung, H., Jäckle, H. and Vorbrüggen, G. (2006). Mutations of the Drosophila zinc finger-encoding gene vielfältig impair mitotic cell divisions and cause improper chromosome segregation. Mol. Biol. Cell 17: 2356-2365. PubMed Citation: 16525017

Struffi, P., Corado, M., Kaplan, L., Yu, D., Rushlow, C. and Small, S. (2011). Combinatorial activation and concentration-dependent repression of the Drosophila even skipped stripe 3+7 enhancer. Development 138(19): 4291-9. PubMed Citation: 21865322

ten Bosch, J. R., Benavides, J. A. and Cline, T. W. (2006). The TAGteam DNA motif controls the timing of Drosophila pre-blastoderm transcription. Development 133: 1967-1977. PubMed Citation: 16624855

date revised: 15 December 2011

Home page: The Interactive Fly © 2008 Thomas Brody, Ph.D.