|Home | About | Journals | Submit | Contact Us | Français|
Many of the key regulators of Drosophila CNS neural identity are expressed in defined temporal orders during neuroblast (NB) lineage development. To begin to understand the structural and functional complexity of enhancers that regulate ordered NB gene expression programs, we have undertaken the mutational analysis of the temporally restricted nerfin-1 NB enhancer. Our previous studies have localized the enhancer to a region just proximal to the nerfin-1 transcription start site. Analysis of this enhancer, using the phylogenetic footprint program EvoPrinter, reveals the presence of multiple sequence blocks that are conserved among drosophilids. cis-Decoder alignments of these conserved sequence blocks (CSBs) has identified shorter elements that are conserved in other Drosophila NB enhancers. Mutagenesis of the enhancer reveals that although each CSB is required for wild-type expression, neither position nor orientation of the CSBs within the enhancer is crucial for enhancer function; removal of less-conserved or non-conserved sequences flanking CSB clusters also does not significantly alter enhancer activity. While all three conserved E-box transcription factor (TF) binding sites (CAGCTG) are required for full function, adding an additional site at different locations within non-conserved sequences interferes with enhancer activity. Of particular note, none of the mutations resulted in ectopic reporter expression outside of the early NB expression window, suggesting that the temporally restricted pattern is defined by transcriptional activators and not by direct DNA binding repressors. Our work also points to an unexpectedly large number of TFs required for optimal enhancer function -- mutant TF analysis has identified at least four that are required for full enhancer regulation.
Drosophila embryonic CNS neuroblast (NB) identity is established by the combinatorial action of multiple factors that include cell-cell signaling, signal transduction and transcriptional regulatory events (reviewed by Skeath and Thor, 2003; Doe, 2008; Sousa-Nunes et al., 2010). Once committed, uniquely fated NBs delaminate from the neuroectoderm arriving in a sub-ectodermal proliferative zone where they initiate neural lineage development by undergoing multiple asymmetric divisions generating functionally diverse neural progeny. During lineage development, NBs undergo temporally ordered transitions in gene expression programs that determine the fate of the neurons and glia born into each temporally defined window (Kambadur et al., 1998; Brody and Odenwald, 2000; Isshiki et al., 2001; reviewed by Brody and Odenwald, 2002; Pearson and Doe, 2004; Jacob et al., 2008). Like Drosophila, mammalian neural development is also characterized by ordered transitions in precursor cell gene expression (Hsiau et al., 2007; Nishi et al., 2009). Understanding the molecular events that regulate these ordered transitions in developmental gene expression programs is a major goal of neurobiology. One approach to elucidating the regulatory networks that control neural lineage development and ultimately neural subtype identity is to functionally dissect the cis-regulatory enhancers that drive temporally restricted NB gene expression programs.
cis-Regulatory modules (enhancers) are genomic DNA fragments (~0.5 to 1.0 kb long) that contain multiple binding sites for sequence specific DNA binding transcription factors (TFs) that collectively control the temporal and spatial expression dynamics of flanking genes (reviewed by Bulger and Groudine, 2010). Recent studies have shown that transcription factors (TFs) with different DNA-binding domains can bind to the same or overlapping DNA sequences, revealing another level of complexity to combinatory regulation of enhancers. For example, it is now known that the canonical DNA-binding site for bHLH TF factors, the CANNTG E-box (Caudy et al., 1988; Singson et al., 1994), has also been shown to be a DNA docking site for the zinc finger transcription factor Snail (Mauhin et al., 1993) and Iroquois homeodomain factors (Billoni et al., 2005).
While DNA sequence alignments between Drosophila melanogaster genes and their orthologs outside the genus are of limited use in identifying enhancers (Hare et al., 2008), the additive evolutionary divergence among multiple Drosophila species is of great utility for identifying functional conserved sequences within cis-regulatory elements (Odenwald et al., 2005; Brody et al., 2007; Patatsenko et al., 2009). For example, all Drosophila NB enhancers characterized thus far contain conserved sequence blocks (CSBs), made up of DNA binding sites for known and as yet unidentified transcriptional regulators (Brody et al., 2008). Our analysis thus far of Drosophila cis-regulatory elements reveals that the structural complexity of their NB enhancers is formidable, with most characterized enhancers containing over 20 CSBs (Brody et al., 2008). Comparative genomic analysis among vertebrates also reveals that many of their enhancers contain multiple CSBs (Hardison, 2000; Levy et al., 2001; Brody et al., 2007; Li et al., 2010). In addition, mutational analysis indicates that small changes in CSBs can alter enhancer cis-regulation (Estes et al., 2008).
To better understand the functional significance of conserved sequences within enhancers and the basis for early temporal NB gene regulation during CNS development, we have studied the role of CSBs in the function of the nerfin-1 NB enhancer. Located 125 bp upstream of the predicted transcriptional start site, the 567 bp nerfin-1 NB enhancer contains 12 CSBs with an average length of 10 bp. Our analysis of nerfin-1 cis-regulation reveals that its NB expression is temporally restricted to the first two of five waves of delaminating NBs (Stivers et al., 2000; Kuzin et al., 2009).
This study examines the function of the nerfin-1 NB enhancer CSBs by mutational analysis, tests their orientation flexibility within the enhancer by sequence rearrangements and identifies TFs that are required for wild-type expression. Multiple putative TF DNA binding sites were found to be required for nerfin-1 NB enhancer regulatory behavior. For example, independent ablations or mutations of conserved sequences that influence TF binding specificity to each of three identical E-box sequences resulted in incomplete enhancer activity. However, DNA sequence rearrangements also revealed that the precise position of these E-box sequences is not crucial for enhancer function. Interestingly, the introduction of an additional E-box at different locations within non-conserved regions interfered with the enhancer's ability to regulate reporter expression. Mutations of CSB sequences that do not correspond to known TF binding sites also affected the stability of reporter expression or triggered ectopic expression in the early delaminating row one NBs. However, none of the mutations resulted in ectopic reporter expression outside early NB expression, nor did we observe misexpression in cells other than NBs, indicating that enhancer regulatory behavior is most likely controlled directly by transcriptional activators and not by direct silencing from DNA binding repressors. A search for TFs that are required for full enhancer activity uncovered four putative regulators. Our analysis reveals that even for this relatively small enhancer, a surprisingly large number of sequence elements are required differentially for its wild-type expressivity in different NBs.
nerfin-1 mRNA is first transcribed during embryonic stages 8 and 9 in early delaminating S1 and S2 ventral cord NBs (Stivers et al., 2000; Kuzin et al., 2009; for ventral cord NB map see Schmid et al., 1999). By embryonic stage 10, expression is also detected in many cephalic lobe NBs. However, during stage 10, expression is significantly down-regulated in ventral cord NBs, and only a small subset of brain NBs contain detectable levels of nerfin-1 transcript by embryonic stage 11. Enhancer-reporter transgene analysis of DNA sequences that span an 11 kb nerfin-1 rescue fragment (Kuzin et al., 2005) located the early NB enhancer, and further deletion analysis of this region revealed that the enhancer is contained within a 576 bp fragment 125 bp upstream of the predicted transcription start site (Kuzin et al., 2009). Similar to the nerfin-1 endogenous mRNA expression, NB enhancer-reporter mRNA transgene analysis revealed that the enhancer drives reporter expression in early delaminating ventral cord NBs during a narrow temporal window of ~2 hours that spans embryonic stages 8 through 11 (Fig. 1A). To identify DNA sequences within the enhancer that are essential for its regulatory behavior, we generated mutant enhancer-reporter transgenes that contained DNA sequence deletions, rearrangements or site-directed base-pair substitutions, that correspond to both known TF DNA-binding sites and novel sequences.
An EvoPrint of the 576 bp nerfin-1 NB enhancer region identified 12 blocks of conserved sequences that were ≥6 bp in length, with the average CSB spanning 10 bp (Fig. 1B). cis-Decoder analysis of these CSBs revealed that they contain multiple DNA sequences that correspond to potential DNA binding sites for known TFs, and many harbor sequences that are also found in CSBs from in vivo characterized NB enhancers associated with the nervy, deadpan, scratch, snail and worniu genes (highlighted sequences in Fig. 1B and C). For example, the nerfin-1 NB enhancer contains three identical E-box TF binding sites that all contain the same core sequence (CAGCTG), and identical sites are also present multiple times in the worniu, deadpan, scratch and nervy NB enhancer CSBs (Fig. 1C), indicating that multiple E-box sites are a common feature of many early temporal network NB enhancers. Other identified TF core DNA binding recognition sites in conserved nerfin-1 NB enhancer sequences include an Antennepedia class homeodomain binding motif (reviewed by Gehring et al., 1994), a core recognition sequence for Nkx-2.5 (Chen and Schwartz, 1995), a core-binding site for Tramtrack-like factors (Fairall et al., 1993) and a consensus binding site for the PRDM1 TF (Roy and Ng, 2004; Hernandez-Lagunas et al., 2005) (see highlighted sequences in Fig. 1B and C).
Previous analysis of the nerfin-1 cis-regulatory DNA delimited the boundaries of the NB enhancer by standard promoter truncation analysis (Kuzin et al., 2009). Here, we extend these studies by examining the functional significance of the enhancer CSBs via site-directed sequence substitutions and small deletions that span individual or groups of CSBs. We designed the DNA mutations so as to avoid generating de novo DNA binding sites for known TFs or sequences that corresponded to conserved elements present in our cis-Decoder tag libraries of previously characterized Drosophila enhancers (Brody et al., 2007). At least three independent transformant lines for each mutant construct were analyzed, to ensure that our expression phenotypes were consistent.
Mutant analysis of the CSBs revealed that each is required for full enhancer function. For example, mutations in the first CSB cluster revealed that this region contains sequence elements that are required for expression fidelity in subsets of NBs or uniformly in intermediate and lateral rows of NBs (Figs. 2, ,3,3, and Supplemental Fig. 6). Deletion of CSBs that contain the conserved CAGCTG E-boxes, along with flanking bases, resulted in a severely diminished or complete abrogation of expression (Fig. 2; #4, #5 and #7). Expression was detected only in two cephalic lobe NBs when the second/middle E-box and flanking sequences were deleted (Fig. 2B; #5 insert), and deletion of the third conserved bHLH binding site and the adjacent CSB resulted in variable un-patterned, stochastic reporter expression in each of the ventral cord segments (Fig. 2B; #7). Deletion #6 (Fig. 2; #6) resulted in loss of expression in intermediate row NBs, while deleting 8 bp from the last CSB resulted in variable/stochastic expression (Fig. 2; #8).
In addition to testing small deletions, we mutated sequence elements within CSBs that are shared with other NB enhancer CSBs, as determined by cis-Decoder analysis (Fig. 1C). Disruption of each of the CSBs in the nerfin-1 NB enhancer by changing two or three base-pairs resulted in either moderately or severely altered expression, and in many cases resulted in stochastic reporter expression (Figs. 3 and and4).4). The variable, un-patterned expression was also apparent among adjacent segments and between corresponding segments of identically staged embryos. Reporter expression phenotypes of five of the 20 base-pair mutations are illustrated in Fig. 3 and the results for the other 15 are shown in Supplemental Fig. 6. Compared to wild-type enhancer-reporter expression (Fig. 3B; wt), the two base-pair substitution #4 resulted in a markedly decreased expression in the lateral 3-5NB (Fig. 3; #4 arrow). In addition, the medial 5-2NB reporter expression varied: it was either undetectable in both hemisegments (Fig. 4B, #4 arrowhead), expressed in only one, or expressed in both. The expression level was also variable in other NBs. The presence of 12 CSBs suggests multiple TFs regulate this enhancer, and the multiplicity of the regulatory effects reflected in the results of our mutant reporter assays also suggests a high combinatorial complexity of interacting factors impacting the NB expression of nerfin-1. These results are comparable to those observed in a similar analysis of the sparkling eye enhancer of the Drosophila shaven gene (Swanson et al., 2010), which interestingly is located on the 4th chromosome and exhibits fewer CSBs than most other developmental enhancers.
Mutations in only one of the CSBs resulted in ectopic reporter expression in NBs, but not outside of the early temporal expression window. The two base-pair substitution #9 resulted in ectopic expression in anterior row 1 NBs of each ventral cord segment, the ventral 1-1NB, the intermediate 3-2NB and the lateral 2-5NB (Fig. 3; #9). Therefore, it appears that the TFs that interact with this CSB function as row specific NB repressors. Mutation of the central two bases of the ATTA homeodomain DNA binding site (Fig. 3B; #10) resulted in an overall reduced expression of the reporter in lateral NBs, including complete absence of expression in several lateral NBs, but with different extents of disruption in each segment. Finally, a triple base-pair mutation in the last CSB (ACT→GCT) resulted in an even more impaired expression (Fig. 3B; #19). There was also a complete loss of detectable reporter expression in many NBs, particularly in row 5 NBs. Our mutagenesis of the conserved sequences revealed that no single mutation affected all NB expression the same way indicating that the environment within each of the early delaminating NBs may differ with regards to the TFs that they express and those that regulate the activity of this enhancer. The complexity of TF environment among NBs is also evident from their lineage development, as each express different developmental programs that result in different uniquely fated neurons and glia (reviewed by Spindler and Hartenstein, 2010 and Sousa-Nunes et al., 2010).
The nerfin-1 NB enhancer contains three conserved CAGCTG E-boxes. Interestingly, cis-Decoder analysis of other NB enhancers revealed that besides nerfin-1, three other in vivo characterized NB enhancers (scrt, wor and dpn) also contain three conserved CAGCTG E-boxes (Brody et al., 2008). To determine if the CAGCTG E-boxes are necessary for nerfin-1 NB expression, we examined nerfin-1 mRNA expression in null mutations for both the ac and sc genes, whose encoded TFs have been shown to bind to CAGCTG E-boxes (Singson et al., 1994). Altered but not complete absence of expression was observed in the embryos homozygous for a chromosomal deficiency that removes both genes (Kuzin et al., 2009). Although the altered nerfin-1 expression indicates that one or both of these bHLH TFs is required for NB expression, the reduced expression may be an indirect result due to the loss of these TFs during earlier phases of NB formation.
To test the requirement for each of the three E-boxes and their specificity, we either disrupted the individual E-box sequences or altered the core central GC bases of each sequence (Fig. 4A). Disruption of each of the bHLH docking sites was accomplished by a two base-pair swap (AG→CT), converting CAGCTG to CCTCTG. We chose the AG→CT substitution because that mutation did not create a sequence element that was present in the conserved CSBs of other enhancers (Brody et al., 2008).
Disruption of each of the E-boxes diminished transgene reporter expression (Fig. 4B). Ablation of the first and second E-boxes triggered altered reporter expression in each of the ventral core hemisegments such that there was a loss of bilateral symmetry in the expression pattern. This resulted in adjacent hemisegments containing different subsets of NBs with different levels of reporter expression. For example, disrupting the first E-box resulted in a lack of bilateral symmetry in NBs expressing the reporter (Fig. 4; #2). Disruption of the third E-box site appeared to variably diminish expression in all NBs, with expression in the lateral NBs being affected to the greatest degree, but nevertheless, keeping intact the overall spatial and temporal wild-type enhancer regulatory behavior (Fig. 4; #8). Taken together, these findings demonstrate a requirement for all three of the E-boxes, and loss of any one of the sites creates an unstable state that results in different levels of enhancer activity in different NBs. Other TFs have been shown to interact with E-box sequences (see Introduction). Therefore multiple E-box-binding factors expressed in different subsets of NBs may be required for nerfin-1 wild-type expression. For example, a member of the Snail family of transcription factors, Worniu, may regulate nerfin-1 NB enhancer via these sequences (Mauhin et al., 1993; see Section 2.6).
We next sought to determine whether an additional CAGCTG E-box would augment or disrupt nerfin-1 NB enhancer expression. To accomplish this, we generated single extra CAGCTG E-box sites at different locations within non-conserved regions of the enhancer (Fig. 4A; #1, #4 and #7). The first was created by mutating 4 bp in the poorly conserved region that is not essential for enhancer function (see section 2.5.) between CSBs 3 and 4 (GTTCTT→CAGCTG, Fig. 4A; #1); the second was created by mutating 3 bp in the non-conserved region between E-box #1 and #2 (CTATTG→CAGCTG, Fig. 4A; #4) and a third was created by inserting 7 bp adjacent to the 3’ E-box (CAGCTGT, Fig. 4A; #7). In each case, the altered enhancers drove reporter expression in a pattern that resembled the basic pattern of wild-type enhancer expression, albeit with overall diminished expression, indicating that an additional CAGCTG E-box did not augment but rather interfered with the reporter expression (Fig. 4).
Previous cis-Decoder analysis of NB enhancers revealed that although many early NB enhancers harbor multiple conserved E-boxes, not all contained the CAGCTG E-box (see Brody et al., 2008; Table 3). For example, the oli gene early NB enhancer has multiple conserved CATGTG E-boxes and no CAGCTG motifs (data not shown). To test the specificity of the core bases of each of the nerfin-1 E-boxes, we altered the central bases by a GC→TG mutation. The core substitutions in each of the E-boxes resulted in markedly diminished transgene reporter activity in the case of the first and third E-box sites, with a less severe disruption in the case of the second site. For the second site, however, reporter expression was almost completely absent in second row lateral NBs (3-5NB; see arrow in Fig. 4B, #6). Therefore, these studies demonstrate that all three specific CAGCTG E-boxes are essential for full enhancer function.
In addition to mutational analysis, we also tested whether the arrangement or orientation of the CSBs was essential for enhancer regulatory behavior. A distinction can be made between a flexible ‘information display’ model of enhancer structure and regulation vs. the more rigidly structured enhanceosome model (reviewed by Kulkarni and Arnosti, 2003). In the former model, enhancer sequences engage in multiple, orientation independent contacts with some or all of the enhancer bound proteins, which in turn independently interact with other co-factors leading to transcriptional regulation. In the more structured model, a complex of DNA binding TFs provide a stereo-specific interface for interaction with the basal transcriptionalmachinery.
To examine whether the orientation and arrangement of CSBs within the cluster is important, we inverted the central enhancer region (Fig. 2A, blue boxed sequence), and found no altered reporter expression when compared to the control wild-type reporter expression (data not shown). To further test the flexibility of CSB ordering, we transposed two regions that contain CSBs #4 and #6 (Fig. 2A; over-lined sequences). Similar to changing the orientation of the CSBs, transposing these two segments did not noticeably affect the NB expression dynamics of the reporter (data not shown). Although we have not rigorously tested all possible combinations of CSB order and orientation, the results indicate that within the resolution of our analysis, rearranging the order and orientation of the nerfin-1 NB enhancer CSBs did not adversely affect enhancer regulatory behavior. We also tested whether DNA sequences within the less- or non-conserved region between CSBs #3 and #4 are essential for enhancer function by generating deletions in the 162 bp less-conserved region (Fig. 2A; deletions #2 and #3). Embryonic reporter expression analysis of the transgenes that contain these deletions did not identify any altered expression when compared to the wild-type enhancer reporter expression (data not show).
Taken together, these rearrangements reveal that there is a certain degree of flexibility in the order and orientation of conserved elements within the nerfin-1 NB enhancer. The absence of detectable effects of alteration of context of CSBs in our study suggests that the display model for enhancer function is preferred over the more rigid enhancesome model. However, our studies do not rule-out the possibility that limited short-range constrained, ordered arrangements between some of the adjacent CSBs are necessary to impart full enhancer function, because studies on other developmental enhancers have detected subtle higher-order sequence requirements (Papatsenko et al., 2009).
As an initial step toward identifying TFs that regulate nerfin-1 NB expression, we examined nerfin-1 mRNA NB expression in different TF loss-of-function mutant backgrounds. Our TF mutant studies have focused on those known TFs that are expressed during early neurogenesis, especially in S1 and S2 NBs (Table 1). Of the 23 different TF mutant alleles examined 10 did not exhibit detectable altered nerfin-1 NB expression and another 9 mutants had altered nerfin-1 expression patterns that were most likely due to loss-of-function during earlier segmental development, since the disrupted NB expression was confined to discrete regions within the ventral cord that overlap the TFs’ earlier role in segmental development (Table 1. and data not shown).
Our expression screen uncovered four TFs that resulted in altered nerfin-1 NB expression throughout the developing ventral cord (Table 1 and Fig. 5). For example, analysis of nerfin-1 expression in the bHLH TF daughterless (da) loss-of-function mutant revealed lower levels of NB expression, with significant reduction in lateral and intermediate NB columns (Fig. 5). Da is ubiquitously expressed in embryos (Cronmiller and Cummings, 1993) and in addition to forming homodimers (Cabrera et al., 1994), it dimerizes with other bHLH TFs that are expressed in subset of NBs (Cabrera and Alonso, 1991). Because of the ability of Da to interact with multiple bHLH factors and also function as a homodimer, it is not surprising that Da by itself or with other bHLH partners may regulate the nerfin-1 NB enhancer. Our previous studies have shown that the loss of both Achaete and Scute bHLH TFs reduces nerfin-1 expression in NBs (Kuzin et al., 2009). However, given that both ac and sc are expressed in subsets of NBs after NB delamination, the reduced expression of nerfin-1 seen in specific loss-of-function mutants may be due to earlier defects in NB determination.
The phenotypic effect in loss-of-function worniu (wor) mutant embryos is similar to that of da in that lower levels of nerfin-1 expression are found in many NBs (Fig. 5). wor is expressed in all early NBs, and mutation has been shown to control NB asymmetric division and cell cycle (Cai et al., 2001; Ashraf and Ip, 2001). We currently do not know whether the worniu effect on nerfin-1 expression is direct or indirect, given that the Wor DNA-binding characteristics have not yet been studied. However, Wor is a member of the Snail zinc-finger transcription factor family and Snail has been shown to bind to E-box sequences (Mauhin et al., 1993).
The strongest reduction in nerfin-1 NB expression observed in our screen was in prospero (pros) loss-of-function mutant embryos. Previous studies of pros expression during neural development have shown that Pros protein is expressed cytoplasmically in all NBs and asymmetrically distributed into GMCs where it is concentrated in the nucleus (Mitsuzaki et al., 1992). The loss of nerfin-1 expression in pros mutant embryos indicates that in addition to its role in GMC gene regulation, Pros also regulates gene expression in the NB as previously shown by (Southall and Brand, 2009). It is unlikely that Pros effect on nerfin-1 expression is through direct binding to the NB enhancer, since no consensus Pros DNA-binding site is present in the NB enhancer, and binding of Pros to the NB enhancer region was not detected in a previous study (Choski, et al. 2006).
Our TF screen also uncovered a member of the Sox family TFs, Dichaete (D), as being required for nerfin-1 expression (Fig. 5). Previous studies have shown that D is required for establishing proper NB identity (Zhao et al., 2007). Since the nerfin-1 NB enhancer does not contain consensus Sox DNA-binding sites (Harley et al., 1994) either in conserved or in non-conserved sequences, it is unlikely that D is a direct regulator of nerfin-1 NB expression.
Loss of hunchback or Kruppel function had no effect on endogenous nerfin-1 expression, suggesting that nerfin-1 regulation via its early NB enhancer is independent of these temporal determinants and that the TFs that regulate early nerfin-1 expression operate outside of the known NB temporal regulatory network. We did observe that prolonged expression of hunchback (using sca-GAL4 driver) triggered an increase in the number of GMCs and nascent neurons expressing nerfin-1 mRNA (data not shown). This observation is consistent with previous studies demonstrating that miss-expression of hunchback generates additional early temporal network progeny (Isshiki et al. 2001). The ectopic expression of Kruppel during NB lineage development (using sca-GAL4 driver) did not noticeably affect nerfin-1 NB expression or expression in GMCs and neurons (data not shown). Similar ectopic expression of prospero (using sca-GAL4 driver) resulted in enhanced nerfin-1 expression in neurons, perhaps reflecting a requirement of prospero for nerfin-1 expression (Kuzin, et al. 2005; Choksi, et al. 2006)
The principle finding of this study is that conserved sequences within the nerfin-1 NB enhancer are essential for its cis-regulatory behavior. Our mutational analysis of the enhancer CSBs, revealing that each is required for wild-type expression, indicates that an unexpected large number of putative interactions between this enhancer and multiple TFs, both activators and repressors, are required for optimal enhancer function. In addition, these studies reveal that many of the mutations had differential affects on enhancer expression within different subsets of NBs. This heterogeneity most likely reflects the different TF combinatorial complexities that exist in each of the uniquely fated NBs. Our studies also demonstrate the importance of specific E-box sequences in the regulation of nerfin-1 NB expression, and suggest that different E-box-binding factors regulate the NB enhancer. These results are consistent with previous studies indicating that ‘proneural’ TFs are required for gene regulation after NB determination (Portman and Emmons, 2000; Seo et al., 2007; reviewed by Hobert, 2005). Our mutational analysis also demonstrated that all three, and only three, CAGCTG E-boxes are required for optimal function, since reporter expression was significantly affected by increasing, decreasing or altering the core E-box sequences. In many cases, our base-pair mutations and CSB deletions resulted in altered un-patterned stochastic reporter expression. Although we do not currently know why certain mutations trigger variable expressivity among NBs, the variability might be due to the creation of a metastable enhancer activation state within individual NBs. For example, the complete or partial loss of one TF binding-site may destabilize that factor's interaction with other co-factors that are required for transcriptional regulation. Identification of essential TFs that directly-bind to the nerfin-1 NB enhancer will provide additional insights into the molecular details of temporally restricted nerfin-1 cis-regulation.
Evolutionarily conserved sequences among eight or more Drosophila species of ≥6 bp within the nerfin-1 NB enhancer were identified by EvoPrinter analysis (Odenwald et al., 2005; Yavatkar et al., 2008). The collective or additive evolutionary divergence of the 8 species used to generate the EvoPrint was ~100 My (Fig. 1B). To identify repeated and unique conserved sequence elements within the nerfin-1 NB enhancer, its CSBs were aligned with each other using the CSB-aligner algorithm (Brody et al., 2008). Conserved sequence elements shared with other NB enhancers were also identified via CSB alignments to the in vivo characterized Drosophila NB enhancers.
Base-pair substitutions within the nerfin-1 NB enhancer CSBs were prepared using previously described techniques (Ho et al., 1989: Horton et al., 1989) using the nerfin-1 genomic rescue fragment as template DNA. Primer sequences used to generate wild-type and mutant enhancer constructs are provided in Supplemental Table 1. Inversion and translocation enhancer constructs were also generated from the rescue fragment using the same procedures. CSB deletion constructs were generated by ligating two PCR fragments for each deletion, replacing each CSB with a Bgl II restriction site. Cloned enhancer variants were sequenced to verify mutations and recloned into pRed H-Stinger (Barolo et al., 2004) or into the modified pCa4B vector (Markstein et al., 2008) as indicated in Supplemental Table 1. The pCa4B vector was modified by inserting into the XbaI-SpeI site the transgene reporter construct from pRed HStinger vector (including its polylinker, minimal HS promoter, RFP ORF, NLS and the SV40 3’ UTR). Cloning details are available upon request. We have not detected any differences in transgene expression between the pRed H-Stinger and the modified pCa4B vector. The cloning steps used to generate the mutant enhancers are illustrated in Supplemental Figs. 1-5. Details of the cloning steps are also available upon request.
Transcription factor loss-of-function mutant allele lines were obtained from the Bloomington Drosophila Stock Center at Indiana University (see Table 1) and from the Chris Doe laboratory. P-element transformants were generated with Df(1)w67c,y or w1118 Drosophila strains using standard techniques based on the methodology described in Rubin and Spradling (1982). For second chromosome site-specific P-element integrations, transformants were generated with the y, w; y+[attp40] strain as previously described (Groth et al., 2004: Markstein et al., 2008) using a modified pCa4B vector (see above). Embryo RFP in situ hybridizations were performed on multiple independent transformant lines for each construct to assure that results were reproducible. sca-Gal4 (Mlodzik, et al. 1990) and pros-Gal4 (Ohshiro, et al. 2000), UAS-hb, UAS-Kr (Grosskortenhaus, et al. 2005) and UAS-prosL17u-2 (Matsuzaki) were obtained from Chris Doe. Standard animal husbandry procedures were used in the care and handling of Drosophila stocks (Ashbruner, 1989).
Embryo collections and fixations were carried out according to the procedures described in Patel (1994). For in situ hybridizations, riboprobes were prepared from a PCR amplified RFP ORF within the pRed H-Stinger vector (Barolo et al., 2004). Roche DIG RNA Labeling Mix protocol was used, and staining was visualized using anti-FITC Fab fragments coupled to alkaline phosphatase. Transgene-reporter expression for each of the enhancer mutations was examined in at least three independent transgene-reporter transformant lines. More detailed protocols for embryo processing and in situ hybridization are available upon request. After whole-mount in situ hybridization, embryos were filleted, viewed in 70% glycerol with 30% phosphate-buffered saline (PBS) and photographed using a Nikon Optiphot microscope equipped with DIC/Nomarski optics. Embryo developmental stages were determined by morphological criteria (Campos-Ortega and Hartenstein, 1985).
Appendix A. Supplementary data
Supplemental Table 1. PCR primers used for wild type and mutant constructs.
Supplemental Fig. 1. A) A linear cartoon of the nerfin-1 NB enhancer with 1 and 3 representing left and right hand regions respectively, and 2 representing the region to be inverted. Primers designated as in Supplemental Table 1. B) The left hand and central region were amplified from the nerfin-1 rescue construct and after annealing was amplified by PCR to produce a single hybrid product. C) After amplifying the right hand region by PCR, the hybrid left/center product was annealed to the right hand region and amplified by PCR to produce the final inverted fragment. D) Cartoon of the final inverted fragment with inverted region marked by arrow.
Supplemental Fig. 2. A) A linear cartoon of the nerfin-1 NB enhancer with region between 1 and 2 (aqua) representing conserved region #4 and between 2 and 3 (orange) representing conserved region #6. Primers designated as in Supplemental Table 1. B) Left hand fragment was amplified so that conserved region #4 was replaced by conserved region #6. The central fragment was amplified using primers that replaced conserved region #6 by conserved region #4 and replacing conserved region #4 by conserved region #6 and after annealing hybrid fragment was amplified by PCR to produce a single product. C) The right hand fragment was amplified so that it contained conserved region #4 and was then annealed to the hybrid left/central fragment to produce the final translocated fragment. D) Cartoon of the final fragments with translocated regions illustrated by arrows.
Supplemental Fig. 3. A) A linear cartoon of the nerfin-1 NB enhancer with region to be deleted marked black. Primers designated as in Supplemental Table 1. B) Left and right hand fragments were produced with primers that deleted the non-conserved region (#3) in Figure 2. The hybrid fragments were annealed and amplified by PCR to produce a single product. C) Cartoon of the final deletion fragment.
Supplemental Fig. 4. A) A linear cartoon of the nerfin-1 NB enhancer with region to be mutated marked aqua and orange (mutant bases marked xx). Primers (#16-#32 and #35-#43) designated as in Supplemental Table 1. B) Left and right hand fragments were produced with primers that containing the altered bases (marked by xx) to produce double base-pair or larger mutations. The hybrid fragments were annealed and amplified by PCR to produce a single product. C) Cartoon of the final deletion fragment.
Supplemental Fig. 5. A) A linear cartoon of the nerfin-1 NB enhancer with region to be deleted marked black. Primers (#5-#6 and #8-# 12) designated as in Supplemental Table 1. Black line represents deleted sequence. B) Left and right hand PCR fragments were produced with primers that replaced a conserved region by BglII site (marked red) and cloned into the TOPO TA vector (Invitrogen). C) After BglII-digestion and fragment ligation a hybrid fragment was created with a substituted BglII site in the conserved region.
Supplemental Fig. 6. DNA mutations within the nerfin-1 NB enhancer conserved sequences alter its cis-regulatory behavior. Shown are embryonic stage 9 NB expression patterns of mutant nerfin-1 NB enhancer CSBs revealed by mRNA in situ hybridization (filleted embryo ventral cords; anterior up). Panel numbers correspond to base-pair substitutions shown in Fig. 3A. Wild-type transgene-reporter expression is shown in Fig. 3B.
The authors would like to thank Jermaine Ross and Antonios Ekatomatis for their technical assistance and the editorial expertise of Judith Brody. We also acknowledge the Bloomington Drosophila Stock Center and Chris Doe for providing mutant fly stocks. This research was supported by the Intramural Research Program of the NIH, NINDS.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.