Mapping the regulatory modules to which transcription factors bind in vivo is a key step toward understanding of global gene expression programs. A chromatin immunoprecipitation (ChIP)-chip strategy has been developed for identifying factor-specific regulatory regions acting in vivo. This method, called the ChIP-enriched in silico targets (ChEST) approach, combines immunoprecipitation of cross-linked protein-DNA complexes (X-ChIP) with in silico prediction of targets and generation of computed DNA microarrays. Use of ChEST in Drosophila is described to identify several previously unknown targets of myocyte enhancer factor 2 (MEF2), a key regulator of myogenic differentiation. The approach was validated by demonstrating that the identified sequences act as enhancers in vivo and are able to drive reporter gene expression specifically in MEF2-positive muscle cells. Presented here, the ChEST strategy was originally designed to identify regulatory modules in Drosophila, but it can be adapted for any sequenced and annotated genome (Junion, 2005).

To predict Dmef2-dependent CRMs, the Drosophila genome was scanned for modules containing one of the three previously described in vivo-acting Dmef2 binding sites. Because muscle differentiation events are controlled by the synergistic action of MADS-box (Mef2 family) and E-box (Twist and MyoD bHLH family) factors modules were sought containing Mef2- and E-box-binding sites. This scanning procedure led to the identification of ~1,243 potential Dmef2-binding CRMs, from which 99 modules were selected, amplified, and spotted to produce a computed Dmef2-CRM array. Three of four previously described Dmef2-CRMs, located in the vicinity of Paramyosin, Act57B, and Dmef2, were identified during genome scanning and selected for spotting as positive control. The fourth in vivo-acting Dmef2-CRM located close to b3-tubulin did not come out from the screen because of the high number of E-box sites used as the scanning criterion. The CRMs located in the vicinity of genes that are not expressed in the mesoderm or genes of unknown expression and function (some of CGs) were rejected. In parallel, X-ChIP was used to isolate DNA fragments to which Dmef2 binds in vivo. ChIP-DNA immunoprecipitated either with Dmef2 antibody or with nonimmune serum was then labeled and used to probe the Dmef2-CRM array. The three previously described in vivo-acting CRMs were found enriched in ChIP-DNA immunoprecipitated with Dmef2 antibody. Importantly, numerous other in silico-identified Dmef2-CRMs were enriched in ChIP-DNA, thus demonstrating the efficacy of the ChEST method (Junion, 2005).

Of the 99 in silico-predicted Dmef2-CRMs, 62 were enriched in the DNA immunoprecipitated with anti-Dmef2 antibody. The CRM-associated Dmef2 targets included genes expressed in all muscle cell types, in which Dmef2 has previously been reported to function. In addition to expected candidates encoding fusion (Lmd, Hibris) and structural (Ket, Pod1) muscle proteins, a large number of CRM-associated Dmef2 target genes coded for TFs and signal transduction proteins. For example, CRMs upstream of Fz2 and within the introns of Ci and Pan indicate a potential role of Dmef2 in transcriptional regulation of genes transducing to the mesodermal cells ectodermal Wg and Hh signals, whereas CRMs within the introns of If and Pka-C3 suggest that by regulating transcription of these genes Dmef2 is involved in the attachment of muscle fibers and in fiber contraction, respectively. In some cases (e.g., Kettin, NetB, N-cad), several Dmef2-dependent CRMs were mapped in the vicinity of the same gene, highlighting the complexity of transcriptional regulation in which Dmef2 is involved. Analysis of the position of CRMs in relation to adjacent genes revealed that the majority of ChIP-enriched modules are located upstream (42%) or within the introns (39%) of target genes. In these two categories, the most frequent positions of Dmef2-CRMs appear between 1 and 5 kb upstream of the gene and within the first intron (Junion, 2005).

To determine whether the ChEST-identified DNA fragments are able to act as regulatory modules in vivo, ten Dmef2-CRMs by reporter gene transgenesis were tested. Nine of 10 CRMs were found to drive reporter gene expression in Dmef2-positive muscle cells. In all transgenic Drosophila lines, lacZ reporter expression at least partially reproduced endogenous gene expression, indicating that the identified CRMs were bona fide enhancers of adjacent genes. For example, the Ket-1, Ket-2, and Ket-3 CRMs laying in the vicinity of the kettin gene, which encodes a giant muscle protein required for the formation and maintenance of normal sarcomere structure, were found to drive lacZ expression in distinct subsets of differentiating body wall muscles. These data indicate that the pan-muscular expression of kettin is regulated in a muscle-type-specific manner, and by multiple Dmef2-binding enhancers. Interestingly, four other analyzed CRMs located within the introns of N-cad and acon and upstream of fz2, sfl, and Meso18E also drive lacZ expression in discrete subsets of somatic muscle precursors. The muscle-type-restricted activity of these modules suggests that both CRM regulators (Dmef2, E-box factors) and their target genes are involved in different aspects of muscle precursors diversification, including muscle fiber shape and axial positioning. Alternatively, the observed muscle-type-specific expression of lacZ may result from the limited size (250-300 bp) of genomic sequences tested. In embryos carrying a Dmef2-dependent CRM found in the intron of If, lacZ expression is detected in a group of ventrolateral muscles. This lacZ pattern correlates with distribution of endogenous If, which accumulates at the extremities of ventrolateral muscle fibers and is required for their correct epidermal insertion. The reduced level of target gene expression in Dmef2 mutant embryos provides an additional support for Dmef2-dependent in vivo regulation of ChEST-identified CRMs (Junion, 2005).

Dorsal-ventral (DV) patterning of the Drosophila embryo is initiated by Dorsal, a sequence-specific transcription factor distributed in a broad nuclear gradient in the precellular embryo. Previous studies have identified as many as 70 protein-coding genes and one microRNA (miRNA) gene that are directly or indirectly regulated by this gradient. A gene regulation network, or circuit diagram, including the functional interconnections among 40 Dorsal (Dl) target genes and 20 associated tissue-specific enhancers, has been determined for the initial stages of gastrulation. This study attempts to extend this analysis by identifying additional DV patterning genes using a recently developed whole-genome tiling array. This analysis led to the identification of another 30 protein-coding genes, including the Drosophila homolog of Idax, an inhibitor of Wnt signaling. In addition, remote 5' exons were identified for at least 10 of the ~100 protein-coding genes that were missed in earlier annotations. As many as nine intergenic uncharacterized transcription units (TUs) were identified, including two that contain known microRNAs, miR-1 and -9a. The potential functions of these recently identified genes are discussed and it is suggested that intronic enhancers are a common feature of the DV gene network (Biemar, 2006).

The Dl nuclear gradient differentially regulates a variety of target genes in a concentration-dependent manner. The gradient generates as many as five different thresholds of gene activity, which define distinct cell types within the presumptive mesoderm, neuroectoderm, and dorsal ectoderm. Total RNA was extracted from embryos produced by three different maternal mutants: pipe^–/pipe^–, Toll^rm9/Toll^rm10, and Toll^10B. pipe^–/pipe^– mutants completely lack Dl nuclear protein and, as a result, overexpress genes that are normally repressed by Dl and restricted to the dorsal ectoderm. For example, the decapentaplegic (dpp) TU is strongly "lit up" by total RNA extracted from pipe^–/pipe^– mutant embryos. The intron-exon structure of the transcribed region is clearly delineated by the hybridization signal, most likely because the processed mRNA sequences are more stable than the intronic sequences present in the primary transcript. There is little or no signal detected with RNAs extracted from Toll^rm9/Toll^rm10 (neuroectoderm) and Toll^10B (mesoderm) mutants. Instead, these other mutants overexpress different subsets of the Dl target genes. For example, Toll^rm9/Toll^rm10 mutants contain low levels of Dl protein in all nuclei in ventral, lateral, and dorsal regions. These low levels are sufficient to activate target genes such as intermediate neuroblasts defective (ind), ventral neuroblasts defective (vnd), rhomboid (rho), and short gastrulation (sog) but insufficient to activate snail (sna). In contrast, Toll^10B mutants overexpress genes (e.g., sna) normally activated by peak levels of the Dl gradient in ventral regions constituting the presumptive mesoderm (Biemar, 2006).

Visual information received from the three types of photoreceptor neurons (R1-R6, R7 and R8) in the fly compound eyes converges to the external part of the medulla neuropil (M1-M6 layers) in a layer-specific fashion: R7 and R8 axons terminate at the M6 and M3 layers, respectively, whereas lamina neurons (L1-L5) relay R1-R6 to multiple medulla layers (M1-M5). During development, R7 and R8 neurons establish layer-specific projections in two separate stages: during the first stage, R7 and R8 axons sequentially target to the R7- and R8-temporary layers, respectively; and at the second stage, R7 and R8 growth cones progress synchronously to their destined layers. Using a set of mutations that delete different afferent subsets or alter R7 connectivity, the mechanism of layer selection was defined. R8, R7 and L1-L5 afferents target to their temporary layers independently, suggesting that afferent-target, but not afferent-afferent, interactions dictate the targeting specificity. N-cadherin is required in the first stage for R7 growth cones to reach and remain in the R7-temporary layer. The Ncad gene contains three pairs of alternatively spliced exons and encodes 12 isoforms. However, expressing a single Ncad isoform in Ncad mutant R7s is sufficient to rescue mistargeting phenotypes. Furthermore, Ncad isoforms mediate promiscuous heterophilic interactions in an in vitro cell-aggregation assay. It is proposed that Ncad isoforms do not form an adhesion code; rather, they provide permissive adhesion between R7 growth cones and their temporary targets (Ting, 2005).

Thus R7 layer-specific targeting occurs in two distinct stages: at the first stage, R7 afferents target to the R7-temporary layer where they remain for about one day until the mid-pupal stage (50% APF) when the R7 growth cones progress synchronously to their final target layer. The two-step target selection has been observed in the vertebrate hippocampus: during embryonic development, entorhinal axons and commissural and associational axons form transient synapses with Cajal-Retzius cells and GABAergic interneurons, respectively, before they synapse onto their postnatal targets, the pyramidal neurons. The two-stage R7 layer selection might serve to coordinate afferent innervation with target development, as in the hippocampus. Alternatively, it might function to reduce the number of potential targets among which R7 growth cones must choose (Ting, 2005).

Using the mutations that delete different afferent subsets or alter R7 connectivity, the mechanism of R7 layer selection has been defined. The genetic cell-ablation results suggest that R8, R7, and L1-L5 afferents target to their temporary layers independently. In addition, the wild-type R8 axons target correctly when the neighboring Ncad or LAR mutant R7s mistarget to the R8-recipient layer. Conversely, the removal of Ncad in single R8s disrupts R8 targeting without affecting the targeting of the neighboring R7s. Thus, the first stage of medulla layer-selection by R8, R7, and L1-L5 afferents probably involves primarily afferent-target interactions. In contrast, R1-R6 growth cone sorting to different lamina cartridges involves both afferent-afferent and afferent-target interactions, even though it required Ncad and LAR (Ting, 2005).

Developmental analyses of single Ncad mutant R7s revealed that Ncad is required for R7 axons to reach and to remain in the R7-temporary layer during the first layer-selection stage. On the basis of its homophilic activity and mutant phenotypes, it has been proposed that Ncad mediates the interaction between the R7 growth cones and the medulla target neurons. The medulla contains over 50 different types of neurons and many express Ncad during development. It is not technologically feasible at the current stage to remove Ncad activity in all or a large number of medulla neurons without affecting the pattern of the optic lobe. Thus, even though it was found that removing Ncad in small patches of medulla neurons does not affect R7 layer selection, the possibility cannot be ruled out that multiple medulla neurons provide redundant N-cadherin-mediated interactions for R7 growth cones in a similar fashion as L1-L5 neurons do for R1-R6 afferents. Alternatively, Ncad might function as a signaling receptor, rather than as a passive adhesive molecule in R7s. Recent studies demonstrated that the cytoplasmic domains of vertebrate classic cadherins can regulate the actin-cytoskeleton via catenins. The Drosophila Ncad contains the two conserved cytoplasmic regions that interact with catenins in vertebrate cadherins. It is conceivable that Ncad may regulate the actin-cytoskeleton in R7 growth cones in response to target-derived cues. Furthermore, although Ncad and LAR share the same adult phenotype, their differential onset of mutant phenotype and double mutant phenotype suggest that they probably regulate different aspects of the first R7 layer-selection stage. Ncad is required for R7 growth cones to initially target to, and remain in the R7-temporary layer throughout the first target-selection stage, while LAR is only required during the later phase (Ting, 2005).

The comparison between the first and second stages of R7 layer selection reveals a glimpse of the underlying mechanism. (1) Targeting to the R7-temporary layer at the first stage appears to be critical for the R7 axons to reach their final destination. Ncad or LAR mutant R7 axons that mistarget to the R8-temporary layer at the first stage, later proceed to terminate incorrectly as well at the R8-recipient layer. (2) In contrast to the initial target selection which follows the axon outgrowth, all R7 and R8 axons enter the second layer-selection stage at approximately the same time, regardless of when they arrive at the medulla. Interestingly, centripetal growth of R1-R6 terminals and synaptogenesis in the lamina coincide with the second stage of R7 layer-selection. It is tempting to speculate that a global signal triggers the initiation of the second stage (Ting, 2005).

In this study, it is reported that the Ncad gene in Drosophila, and probably in other insects, undergoes alternative splicing to generate multiple isoforms. However, the lack of isoform-specificity, revealed by the transgene rescue, overexpression experiments, and heterophilic interaction assays, argues against the hypothesis that the Ncad isoforms constitute an adhesion code to direct targeting specificity. Instead, the idea is favored that Ncad plays a permissive role in R7 layer selection. Nevertheless, the remarkable conservation of the Ncad alternative splicing over 250 million years of evolution suggests an adaptive advantage for Ncad molecular diversity, whose function awaits further investigation (Ting, 2005).

Ommatidial rotation is a cell motility read-out of planar cell polarity (PCP) signaling in the Drosophila eye. Although the signaling aspects of PCP establishment are beginning to be unraveled, the mechanistic aspects of the associated ommatidial rotation process remain unknown. This study demonstrates that the Drosophila DE- and DN-cadherins have opposing effects on rotation. DE-cadherin promotes rotation; DE-cad mutant ommatidia rotate less than wild type or not at all. By contrast, the two DN-cadherins act to restrict this movement, with ommatidia rotating too fast in the mutants. The opposing effects of DE- and DN-cadherins result in a coordinated cellular movement, enabling ommatidia of the same stage to rotate simultaneously. Genetic interactions, phenotypic analysis and localization studies indicate that EGF-receptor and Frizzled-PCP signaling feed into the regulation of cadherin activity and localization in this context. Thus, DE- and DN-cadherins integrate inputs from at least two signaling pathways, resulting in a coordinated cell movement (Mirkovic, 2006).

Although the role for DE-cad in tissues undergoing rearrangements during development is established, a direct role for DE-cad in cell and tissue movement has been more difficult to study in vivo owing to its essential role in maintenance of epithelial integrity. Analysis of adult eye phenotypes of a homozygous viable shg/DE-cad allele and a dominant-negative DE-cad construct (DE-cadDN), expressed in the R3/R4 and later R1/R6, R7 precursors, indicate that DE-cad is required throughout the rotation process. The ability of ommatidia to complete the precise 90° rotation directly depends on DE-cad activity. Both the extracellular domain, responsible for cell-cell adhesion, and the intracellular domain, linking DE-cad to the actin cytoskeleton, are required for rotation. DE-cad associates with the actin cytoskeleton primarily through interactions with Arm/ß-catenin. Although ß-catenin has a dual role in cell adhesion and Wg signaling (which can be separated), these data indicate that during ommatidial rotation ß-catenin acts through its role in cell adhesion (Mirkovic, 2006).

Ommatidial rotation represents the final step in establishing PCP during eye development. The direction of rotation depends on proper R3/R4 cell fate specification, which is determined by PCP signaling. The Egfr pathway and input by rotation-specific genes, e.g. nemo, are thought to function in parallel to Fz-PCP signaling. An enhancement of the sev>DE-cadDN rotation defects was observed by dose reduction in core regulatory PCP genes dgo and stbm; ommatidial under-rotation and the number of ommatidia that did not initiate rotation in sev>DE-cadDN/dgo^-/+, stbm^-/+ was comparable with the enhancement of sev>DE-cadDN by heterozygosity of a shg null allele). The localization of PCP protein complexes at the level of adherens junctions is consistent with the idea that PCP factors can influence DE-cad function. The mechanism of this regulation remains unclear. The RhoA-RNAi transgene, which was expressed only in R3/R4 precursors during the initiation of ommatidial rotation, enhanced sev>DE-cadDN associated under-rotation defects. Although a RhoA requirement in multiple cellular processes makes it difficult to dissect its specific role in rotation, the specificity of the phenotype (enhanced under-rotation in sev>DEcad/RhoA^IR) suggests a role for RhoA in the regulation of cadherin-mediated cell movement (Mirkovic, 2006).

Although Egfr signaling appears to be required for the precise 90° rotation, its role in the process - promoting motility or antagonizing it - has remained unclear. The genetic data suggest that Egfr signaling acts positively to promote rotation, since a reduction in Egfr signaling enhances the sev>DE-cadDN under-rotation phenotype. This may reflect a positive role for Egfr signaling in the regulation of DE-cad activity or turnover at the membrane, as suggested from human tumor cell lines. Affecting the function of endocytic pathway components can also have an effect on ommatidial rotation. This might be mediated by Egfr signaling, as is thought to be the case in human cancer cells, leading to recycling and redistribution of E-cad at the plasma membrane (Mirkovic, 2006).

Drosophila DN-cadherins, which are encoded by the adjacent cadN and cadN2 genes, are the main cadherins expressed in the nervous system. In developing photoreceptors they participate in axon guidance, and in pupal eye discs they mediate terminal patterning of the retina [through specific expression in cone cells. During PCP establishment, DN-cad1 is concentrated at the border between R3/R4 precursors, in a pattern largely complementary to DE-cad. This suggested a possible combinatorial role for DE-cad and DN-cad in rotation, with DN-cad either providing a structural role in rotating clusters, or participating in signaling cascades that regulate cell movement. Analysis of DN-cad mutant clones in discs during rotation demonstrated a specific function; many mutant clusters have completed rotation well before wild-type clusters of the same stage. These data indicate that DN-cadherins function to slow down rotation, serving an opposing function to DE-cad (Mirkovic, 2006).

The balance and complementary distribution of DE-cad and DN-cad appear crucial for correct rotation to occur. Mild overexpression of DN-cad1 in R3/R4 (sev>DN-cad) is sufficient to interfere with the process, possibly by affecting DE-cad levels. Consistently, DN-cad1 overexpression enhances sev>DE-cadDN induced under-rotation and overexpression clones of DN-cad1 cause a decrease in endogenous DE-cad levels. Alternatively, the negative effect of DN-cad on DE-cad might be through competition for ß-catenin, since sev>DE-cadDN is partially rescued by UAS-Arm^S2, although since sev>DN-cad is not enhanced by arm dose reduction this appears less likely. Interestingly, sev>DN-cad is enhanced by co-expression of full-length DE-cad and full-length Arm. These phenotypes resemble those of a strong sev>DN-cad line, suggesting that DN-cad is stabilized by increased levels of available Arm, and also that co-overexpression of two cadherins may interfere with optimal turnover rate at the membrane (Mirkovic, 2006).

The cadherin superfamily of cell adhesion molecules have been proposed to play important roles in determining synaptic specificity in developing nervous systems. Function was studied of N-cadherin in Drosophila second order olfactory projection neurons (PNs), each of which must selectively target their dendrites to one of 50 glomeruli. The results do not support an instructive role for N-cadherin in selecting dendritic targets; rather, N-cadherin is essential for PNs to restrict their dendrites to single glomeruli. Mosaic analyses suggest that N-cadherin mediates dendro-dendritic interactions between PNs and thus contributes to refinement of PN dendrites to single glomeruli. N-cadherin is also essential for the development of PN axon terminal arbors in two distinct central targets: regulating branch stability in the lateral horn and restricting high-order branching in the mushroom body. Although the N-cadherin locus potentially encodes eight alternatively spliced isoforms, transgenic expression of one isoform is sufficient to rescue all phenotypes (Zhu, 2004).

The olfactory neural circuit presents a fascinating wiring problem. From flies to mammals, olfactory receptor neurons (ORNs) expressing a common olfactory receptor have widely distributed cell body positions in sensory epithelia, yet their axons converge onto the same glomeruli in the antennal lobe/olfactory bulb, thereby forming the first spatial odor map in the brain. In Drosophila, second order olfactory projection neurons (PNs), the main postsynaptic targets of ORNs, send dendrites into specific glomeruli according to their lineage and birth order. Furthermore, PNs of a particular glomerular class exhibit stereotypical axon branching patterns and terminal fields in the lateral horn, one of the higher order olfactory centers, allowing stereotypical transfer and transformation of olfactory information farther into the brain. Thus, during the construction of the fly olfactory system, a given ORN must choose one of 50 glomeruli to target its axons, while a given PN must also choose one of 50 glomeruli to target its dendrites, and furthermore the PN must coordinate its dendritic target choice with its axon terminal arborization pattern in higher olfactory centers. Evidence suggests that at least the initial development of the olfactory circuit is independent of sensory input; for instance, the antennal lobe glomerular assembly, as well as the axon terminal arborization pattern, are largely complete before the first signs of olfactory receptor expression. Thus, wiring seems to be controlled predominantly by genetic programs. The molecular logic that underlies this striking wiring specificity is largely unknown (Zhu, 2004).

The wiring specificity of the fly olfactory system has been studied from the perspective of the second order PNs. Because lineage and birth order provide important information to determine where a PN should send its dendrites, it is likely that PNs of different glomerular classes express different combinations of transcription factors. A specific set of transcription factors might then direct the expression of a unique combination of cell surface receptors within PNs of a particular glomerular class, allowing their dendrites to choose a specific location in the antennal lobe and their axons to exhibit a specific terminal arborization pattern in higher olfactory centers. Indeed, a pair of POU domain transcription factors, Acj6 and Drifter, are differentially expressed in the two main PN lineages and regulate wiring specificity of PNs of these two lineages (Zhu, 2004).

To study cell-autonomous functions of candidate genes for their roles in PN wiring specificity, the MARCM strategy was used to generate positively labeled homozygous mutant clones in an otherwise unlabeled and largely heterozygous genetic background. Analysis focused on projection neurons that express GH146-GAL4, which accounts for 90 of the estimated 150-200 PNs. The majority of GH146-positive PNs (from here on referred to as PNs) are of the anterodorsal (adPNs) and lateral (lPNs) neuroblast lineages that innervate stereotypical, intercalated, yet nonoverlapping sets of glomeruli. These PNs can be visualized as neuroblast clones or single-cell clones using MARCM. Strong loss-of-function mutants in a number of cadherin superfamily cell adhesion molecules including N-cadherin, E-cadherin, Flamingo, Fat, and Dachsous. The only mutant that shows a detectable phenotype in PN dendrite or axon patterning is N-cadherin. The allele N-cadherinM19, which appears to be a complete loss-of-function allele, was used: it contains a stop codon in the extracellular domain, is nearly protein null when assayed with an antibody against an extracellular epitope before the stop codon, and it behaves genetically as a null in embryonic nervous system patterning (Zhu, 2004).

Developmental studies reveal that the dendrites of N-cadherin mutant PNs occupy the same positions as their wild-type counterparts during early pupal development, arguing against a function for N-cadherin in initial dendritic targeting. Furthermore, N-cadherin^−/− dorsal and lateral neuroblast clones largely preserves their dendritic innervation pattern. It remains possible that N-cadherin may function in determining individual PN glomerular choices within either the dorsal or lateral neuroblast lineage. However, analysis of single-cell clones of DL1 and DA1 class (of which independent criteria are available to verify their identities) argues against this possibility. In contrast, the dendritic overspill phenotype is observed in every N-cadherin^−/− single-cell clone, supporting a permissive role of N-cadherin in determining connection specificity in the antennal lobe. Analysis of the role of N-cadherin in ORNs also does not support a major instructive role in target selection of ORN axons (Zhu, 2004).

The fact that the N-cadherin locus encodes eight alternatively spliced isoforms makes it attractive to suppose that this molecular diversity is used for selective synaptic adhesion, thereby contributing to synaptic specificity. Yet it has been shown that UAS-GAL4-mediated transgene expression of a single isoform is sufficient to rescue dendritic overspill phenotypes of all classes of PNs examined. This observation suggests that differential expression of N-cadherin at the level of transcription or alternative splicing is not required for PN dendritic targeting specificity (Zhu, 2004).

At present the possibility exists that N-cadherin plays an instructive role in determining targeting specificity of a subset of PNs that were not analyzed (for instance GH146-negative PNs). It is also possible that N-cadherin may play a role in connection specificity in higher olfactory centers, especially in the lateral horn (Zhu, 2004).

This study nevertheless has uncovered an important and novel function for N-cadherin in restricting dendritic targeting to single glomeruli. Every PN class examined exhibited a remarkably similar phenotype: instead of dendrites of a single PN targeting to a single glomerulus, dendrites from N-cadherin mutant PNs spread to neighboring glomeruli. Developmental studies revealed a major function of N-cadherin in dendritic refinement. In wild-type animals, dendrites from a single PN initially occupy a proportionally larger area of the antennal lobe, presumably overlapping with dendrites of other PN classes. They then refine their dendritic domain to single glomeruli by 50 hr APF, concomitant with glomerular maturation. N-cadherin mutant dendrites initially occupy dendritic domains of roughly similar size, but they then fail to restrict their domains to a single glomerulus (Zhu, 2004).

Dendritic refinement of second order olfactory neurons appears to be a common feature in both flies and mammals. Mammalian mitral/tufted cells, equivalent to PNs, also send their apical dendrites into several glomerular structures during early development, subsequently refining to a single glomerulus. The mechanism of this refinement is unclear. Interestingly, loss of odor-evoked neuronal activity of ORNs does not affect mitral cell dendritic refinement. Likewise, Drosophila PN dendritic refinement occurs before any OR receptor expression and before morphologically mature synapses are detected in the antennal lobe, suggesting an activity-independent process (Zhu, 2004).

What kind of cell-cell interactions does N-cadherin mediate during PN dendritic refinement? Previously, N-cadherin has been proposed to mediate axonal target selection in the visual system of Drosophila. During development, PN dendritic refinement coincides with ORN axonal invasion into the antennal lobe; N-cadherin is expressed in both ORNs and PNs at this time. Based on these observations, it is attractive to hypothesize that N-cadherin mediates homophilic interactions between ORN axons and PN dendrites to refine neural processes. However, no positive evidence was found for N-cadherin-mediated interaction between ORN axons and PN dendrites in PN dendritic refinement. Rather, the reverse MARCM analysis demonstrates that N-cadherin in other PNs of the same lineage, including PNs that occupy the same glomerulus, is required for the uniglomerular targeting of wild-type PNs (Zhu, 2004).

What cellular mechanisms could explain both the autonomous and nonautonomous effects of N-cadherin in uniglomerular targeting? The simplest model is that N-cadherin expressed on the surface of PN dendrites confers proper adhesiveness to the dendrite during and after the initial targeting event. Loss of N-cadherin results in reduced cell adhesion, allowing dendrites to more easily invade the neighboring glomeruli or shooting out of the confines of the antennal lobe. The non-cell-autonomous effect of N-cadherin revealed by the reverse MARCM analysis suggests that their interaction partners are dendrites of other PNs, most likely of the same class and thereby occupying the same general domain. In this scenario, direct PN dendrite-dendrite interaction mediated by N-cadherin contributes to confining dendrites to individual glomeruli (Zhu, 2004).

Other alternative scenarios are less likely. For instance, the reverse MARCM phenotype could be explained if altered PN patterning as a result of the unlabeled mutant neuroblast clones leads to altered ORN axons, which then indirectly affect the labeled wild-type PN. However, at least at the VA1lm glomerulus, it has been shown that altered PN dendritic patterning does not result in a corresponding alteration of ORN axon arborization (Zhu, 2004).

Since N-cadherin is expressed in all PN classes, it must work together with other cell surface proteins eventually to refine dendrites to their class-specific glomeruli. Some of these cell surface proteins may allow dendrites to be targeted to their appropriate spatial locations; others may mediate repulsive interactions among dendrites of different PN classes (Zhu, 2004).

Based primarily on the developmental timing of PN dendritic patterning and ORN axon arrival, as well as studies of cellular constituents of the developing antennal lobe, it has been proposed that PN-PN interactions are utilized in creating a prototypic dendritic map before the arrival of ORN axons. Furthermore, since functional studies from mammals to moths suggest an autonomous role for ORNs in organizing glomerular pattern, it is proposed that ORN axons and PN dendrites may be capable of generating independent maps; ORN-PN interaction would then enable the registration of the two maps with one another. Analysis of the N-cadherin mutant phenotype and the interaction between mutant ORNs or PNs lends support to these hypotheses. (1) It is shown that PN-PN interactions do occur and are essential at least for the uniglomerular targeting of dendrites. (2) The fact that dendritic overspill of PNs does not result in a corresponding ORN axon overspill supports the view that ORNs deploy targeting strategies independent of their future postsynaptic partners. At the same time, disruption of N-cadherin in ORNs does not affect the initial dendritic refinement of those neurons. It is possible that further ORN-PN interaction during the synapse maturation period could play a stabilizing role in the maintenance of synaptic connections, possibly explaining why disruption of N-cadherin in ORNs results in a more diffuse dendritic projection in the adult (Zhu, 2004).

This analysis has revealed that N-cadherin has diverse functions in a single type of neuron, olfactory projection neurons. In addition to its important roles in restricting the dendrites of each PN to a single glomerulus in the antennal lobe, N-cadherin is required for PN axon terminal arborization in both the mushroom body and the lateral horn. Do these diverse functions of N-cadherin reflect local actions of N-cadherin in these three sites, or could N-cadherin directly act in one site, with the phenotypes in other sites being secondary consequences due to intraneuronal communication? Several lines of evidence support the notion that N-cadherin acts locally and independently in PN dendrites and axons: (1) N-cadherin is highly expressed in all three sites; (2) based on developmental analysis of the mutant phenotype, N-cadherin is required in overlapping time windows in these three sites. For instance, the antennal lobe phenotype suggests a continuous role of N-cadherin from before 18 hr APF to close to 50 hr APF, whereas its function in the lateral horn for DL1 PNs appears to be at 36-42 hr APF. The lack of clear-cut sequential actions makes the intraneuronal communication model less likely. (3) N-cadherin affects the dendrites of all classes of PNs but the axons of only a subset of PNs. For instance, although N-cadherin mutant DA1 single-cell clones exhibited dendritic phenotypes, the highly specific targeting of lateral horn terminal arborization to the anterior corner for the DA1 PN was unaffected in the N-cadherin mutant. Such uncoupling argues against the intraneuronal signaling model (Zhu, 2004).

Interestingly, axonal phenotypes in the two higher order olfactory centers appear qualitatively different. In the lateral horn, where PN axon terminal arborization is highly stereotyped according to glomerular class, a strikingly specific perturbation of selective terminal branches is observed in two classes of PNs. Developmental studies clearly demonstrate that N-cadherin affects lateral horn axon terminal branch stability, rather than its initial growth. This represents the first demonstration of selective stabilization of terminal branches mediated by a cadherin, which may well be important in establishing the specificity of synaptic connection of PN axons with their postsynaptic targets. Because of the lack of knowledge about the postsynaptic neurons and their dendritic fields in the lateral horn, the significance of loss of particular branches in synaptic specificity shall await future investigations. Finally, it was previously shown that the DL1 dorsal axon branch was also specifically disrupted in PNs deprived of the POU transcription factor Acj6. Interestingly, these ultimately similar phenotypes are mechanistically distinct: Acj6 affects dorsal branch formation, whereas N-cadherin affects its stabilization (Zhu, 2004).

The axonal branching pattern of PNs in the mushroom body is considerably less stereotyped than that in the lateral horn in wild-type animals. Here, at least for one class of PNs (DL1), N-cadherin functions to restrict exuberant higher-order branches, in contrast to its role in stabilizing selected branches in the lateral horn for the same class of PNs (Zhu, 2004).

Drosophila olfactory receptor neurons (ORNs) elaborate a precise internal representation of the external olfactory world in the antennal lobe (AL), a structure analagous to the vertebrate olfactory bulb. ORNs expressing the same odorant receptor innervate common targets in a highly organized neuropilar structure inside the AL, the glomerulus. During normal development, ORNs target to specific regions of the AL and segregate into subclass-specific aggregates called protoglomeruli prior to extensive intermingling with target dendrites to form mature glomeruli. Using a panel of ORN subclass-specific markers, it has been demonstrated that in the adult AL, N-cadherin (N-cad) mutant ORN terminals remain segregated from dendrites of target neurons. N-cad plays a crucial role in protoglomerulus formation but is largely dispensible for targeting to the appropriate region of the AL. It is proposed that N-cad, a homophilic cell adhesion molecule, acts in a permissive fashion to promote subclass-specific sorting of ORN axon terminals into protoglomeruli (Hummel, 2004).

A striking feature of olfactory system organization is the evolutionarily conserved arrangement of ORN terminals into an odortopic map. Here, ORNs expressing the same odorant receptors innervate common targets in a highly organized neuropilar structure, the glomerulus. Mosaic animals in which N-cad was selectively removed from ORNs are largely devoid of glomeruli. ORN targeting to the appropriate region of the AL was not dependent upon N-cad. Developmental analysis reveals that N-cad is essential for protoglomerulus formation at an early stage of AL development. The analysis of N-cad thus supports the notion that targeting and glomerulus formation are distinct steps in constructing an olfactory sensory map (Hummel, 2004).

A series of studies in both vertebrate and invertebrates has underscored the key role played by ORNs in regulating glomerulus formation. Surgical ablation of the antennal primordium from the moth Manduca sexta results in the loss of glomerular structures, as does the genetic disruption of ORN development in Drosophila. Mouse ORNs form glomerular structures when forced to innervate ectopic sites in response to surgical or genetic ablation of the olfactory bulb. These interactions may be instructive, since the pattern of sexually dimorphic glomeruli in gynandromorphs in the moth reflects the sex of the ORNs, not the sex of AL cells. The molecular mechanisms that mediate the intrinsic capacity of ORNs to sort out in a subclass-specific manner are unknown. The finding that genetically mosaic animals in which ORNs deficient in N-cad lack glomeruli provides further evidence that ORNs are essential for glomerular development and simultaneously identify a molecular component required in this process (Hummel, 2004).

Developmental studies revealed a stereotyped pattern of glomerulus development in flies. ORNs send axons into the AL from early (18 hr) through late (100 hr) pupal development. Axons extend into the nerve fiber layer surrounding the incipient AL. Most ORNs project through the commissure to the contralateral AL. Studies with a panneuronal marker reveal that ORNs extend processes from the nerve fiber layer into the developing dendritic layer immediately after they reach the AL, on their way to the commissure. Using the 72OK marker (which labels axons that converge onto two ventral-medial glomeruli, VM1 and VM4), it has been demonstrated that axons of the same ORN subclass extend thin processes into a restricted area of the AL corresponding to the approximate position of the future glomerulus. These processes quickly grow and rapidly condense into a protoglomerulus (Hummel, 2004).

Protoglomeruli are largely peripheral to the mass of dendrites of AL projection neurons. Glomerulus formation ensues as ORN processes and PN dendrites intermingle. The sequence of events leading to glomerulus formation is in many ways phylogenetically conserved. It is important to note, however, that the dynamics of glomerulus formation differ between different subclasses of mouse ORNs. Nevertheless, in both vertebrates and invertebrates, the formation of protoglomeruli largely comprising the processes of ORNs occurs prior to extensive intercalation with dendrites of target neurons (Hummel, 2004).

N-cad mutant ORNs do not form protoglomeruli. 72OK ORNs lacking N-cad extend processes into the appropriate region of the lobe. In striking contrast to wild-type, however, rather than condensing into protoglomeruli at 35% pupal development (PD), ORN axons extend farther throughout the ventral medial region of the AL before retracting back to the surface of the neuropil. They remain segregated from the dendrites of target neurons. While markers for other subsets of ORNs early in development are not available, it is likely that the process visualized with 72OK will extend to other ORNs, as developmental analysis with a panneuronal marker reveals a widespread failure of protoglomerulus formation. Furthermore, adult N-cad mutant ORN processes terminate on the surface of the lobe in close proximity to the position of the glomerulus in a wild-type animal. These studies establish N-cad as a key regulator of the cellular interactions underlying the formation of protoglomeruli (Hummel, 2004).

The subclass-specific convergence of ORN axons reflects specific interactions between ORN terminals within the target region. In both the mouse and fly, the cell bodies of different ORN subclasses are scattered throughout the nasal epithelium and the antenna, respectively. ORN axons do not segregate into subclass-specific fascicles as they project to the target. Association between axon terminals takes place within the target region itself. That different ORNs show affinity for fibers of the same subclass is supported by studies in both organisms. In the mouse extratoes mutant, the olfactory bulb fails to form. ORNs terminate instead in a fibrocellular mass lacking PNs and only a small number of LNs. Despite the lack of target neurons, P2 neurons sort out into a glomerular-like structure. ORNs expressing the same odorant receptors show selective affinity. Mice expressing odorant receptor transgenes in ORNs that lie in inappropriate epithelial zones form ectopic glomeruli in regions of the olfactory bulb distinct from those formed by ORNs expressing the endogenous receptor. In some cases, however, axons from ORNs expressing the endogenous receptors are recruited into these ectopic glomeruli. A similar conclusion has been reached in Drosophila, exploring targeting defects in ORNs mutant for Dscam. ORNs lacking Dscam target to inappropriate regions of the AL; indeed, several subclasses of ORNs terminate in the wrong ganglion in these mutants. Nevertheless, ORNs of the same subclass sort out from fibers of other ORN classes to form separate, often adjacent, ectopic glomeruli (Hummel, 2004 and references therein).

The homophilic cell adhesion activity of N-cad supports a simple model for its function in the developing fly olfactory system. Here N-cad acts to promote interactions between ORN terminals of the same subclass. It is likely that this interaction is largely permissive in nature because N-cad is widely expressed, and removal of N-cad from ORNs leads to a similar axonal phenotype for all 10 ORN subclasses tested. In contrast to the permissive function for N-cad in protoglomerulus formation in flies, odorant receptors appear to play an instructive role in this process in vertebrates. Genetic and expression studies indicate that odorant receptors do not play a role in ORN convergence in flies, although it remains possible that a subset of these receptors not yet tested will, as in vertebrates, serve this function (Hummel, 2004).

Could N-cad play an instructive role in sorting? As ORNs target to the appropriate region of the lobe, the sorting problem is reduced to sorting among a small number of different ORN subclasses within a local region. Modulation of N-cad activity in different classes may provide sufficient specificity for sorting. Indeed, cells in culture expressing different levels of cadherins sort out from one another. As panneuronal expression of a single form of N-cad in ORNs rescues the N-cad mutant targeting phenotypes, it is unlikely that different levels of N-cad protein specified at the level of transcription or different isoforms produced by alternative splicing underlie specific sorting. Presumably, other cell surface components in fly ORNs mediate subclass-specific sorting to protoglomeruli modulating the core homophilic cell adhesion activity of N-cad (Hummel, 2004).

N-cad may not mediate adhesion between ORNs, but rather may promote interactions between ORNs and target dendrites. In the fly visual system, for instance, N-cad is required in R7 photoreceptor neurons to select the appropriate target region. Genetic mosaic studies support the view that this reflects interactions between R7 growth cones and their targets rather than between R7 neurons and other photoreceptor neurons. Since N-cad is expressed on both ORNs and target neurons as ORNs enter the target region, it is expressed at the right place and time to mediate interactions between them. However, recent studies (Zhu, 2004) indicate that glomerulus formation does not require N-cad in PN dendrites. N-cad may also play a subsequent role in mediating interactions between ORNs and targets, but in order to critically address this issue, it will be necessary to assess glomerulus development in animals in which N-cad is removed after formation of protoglomeruli (Hummel, 2004).

While adhesion between ORN terminals has emerged as a common theme from work in both vertebrate and invertebrate systems, specific interactions between axons within the target region may be a more widespread phenomenon. Indeed, in the Drosophila visual system, the exquisite target specificity of different subclasses of R1-R6 photoreceptor axons relies upon interactions between them. In wild-type animals, R1-R6 axons from a single ommatidium project away from each other to a distinct set of targets. This requires N-cad and it has been proposed that N-cad mediates adhesive interactions between R cells and their targets. While biochemical studies support an adhesive function for classical cadherins, the N-cad phenotypes in R1-R6 and ORNs are also consistent with a repellent function. N-cad on R cell axons may promote repulsive interactions between them rather than mediating adhesion between R cell axons and target cells. In the olfactory system, N-cad may promote repulsive interactions between different classes of ORNs, thereby leading to their segregation rather than promoting interactions between axon terminals of the same class. That N-cad may mediate repellent interactions between neurites is supported by genetic analysis by Zhu (2004) of a requirement for N-cad in mediating interactions between PN dendrites. It has been proposed that repellent interactions between R8 axons play a crucial role in elaborating a precise topographic map in the fly visual system and that the protocadherin Flamingo serves this function. Whether interactions between afferents are adhesive or repulsive, or indeed whether they interact in a more complex and instructive fashion, these studies raise the important issue that afferent interactions may play crucial and hitherto poorly appreciated roles in the elaboration of specific patterns of synaptic connectivity (Hummel, 2004).

The molecular mechanisms underlying the formation of an odortopic map presents a complex problem in cellular recognition. A major advance in this field was the discovery that odorant receptors in the mouse play an instructive role in ORN targeting. Given the striking conservation in the cellular organization in the olfactory system of the mouse and fly, it is surprising that fly odorant receptors do not serve this key targeting function. Target specificity in this system must rely on other cellular recognition mechanisms. The Ig superfamily protein Dscam plays a crucial role in targeting some, but not all, ORNs. In the absence of Dscam, some subclasses of ORNs target to inappropriate regions of the AL, where they form ectopic glomeruli. Hence, Dscam contributes to target specificity, whereas N-cad acts at a later step to promote the formation of protoglomeruli. It is anticipated that further studies on N-cad and Dscam, as well as other cell surface and signaling proteins identified in ongoing genetic screens for targeting mutants in the olfactory system, will lead to a detailed molecular understanding of the cellular interactions underlying the exquisite specificity of neuronal connections in the fly olfactory system (Hummel, 2004).

Drosophila N-cadherin is required for the formation of precise patterns of connections in the fly brain. Alternative splicing is predicted to give rise to 12 N-cadherin isoforms. An N-cadherin allele, N-cad^18Astop, was identified that eliminates the six isoforms containing alternative exon 18A. This allele strongly disrupts the connections of R7 photoreceptor neurons. During the first half of pupal development, N-cadherin is required for R7 growth cones to terminate within a temporary target layer in the medulla. N-cadherin isoforms containing exon 18B are sufficient for this initial targeting. By contrast, 18A isoforms are preferentially expressed in R7 during the second half of pupal development and are necessary for R7 to terminate in the appropriate synaptic layer in the medulla neuropil. Transgene rescue experiments suggest that differences in isoform expression, rather than biochemical differences between isoforms, underlie the 18A isoform requirement in R7 neurons (Nern, 2005).