Nervous systems exhibit highly reproducible patterns of connectivity that are essential for their proper functions. Axon pathfinding in many systems is heavily dependent upon genetically programmed expression of guidance factors and their receptors . Recent studies have indicated that dendrite target selection and aspects of synapse specificity can also be precisely genetically programmed in flies  and vertebrates . For example, wiring specificity in the adult Drosophila olfactory system is achieved during pupal development before the onset of olfactory receptor expression . Olfactory receptor neurons (ORNs) project their axons to glomeruli in the antennal lobe (AL), where they synapse with the dendrites of projection neurons (PNs) (Figure 1a; reviewed in [5,6]). PNs target their dendrites to single glomeruli and send their axons to stereotypic locations in the mushroom body (MB) and lateral horn (LH) according to their glomerular class [7-9]. Most PNs are derived from three neuroblast lineages, anterodorsal (adPNs), lateral (lPNs) and ventral (vPNs). adPNs and lPNs innervate intercalating but non-overlapping sets of glomeruli, suggesting lineage-specific control of targeting. Additionally, adPNs are specified by birth order, suggesting that instructive information within a lineage determines wiring patterns . Indeed, PN dendritic patterning precedes ORN axon patterning: by the time pioneering ORN axons arrive at the developing AL, PN dendrites have already formed a coarse map by virtue of their specific dendritic targeting .
Several cell-surface proteins, including Sema-1a, Dscam and N-Cadherin, have been shown to play different roles in PN dendritic development. These studies suggest a model in which PN dendrites first target to a rough region of the AL based on molecular gradients and then are further refined by dendro-dendritic and dendro-axonal interactions [10-12]. The expression of these and additional cell-surface proteins are likely controlled by a transcriptional code that acts to uniquely specify the wiring aptitude of individual PNs[13,14]. Studies of several transcription factors (TFs) support the existence of a transcriptional hierarchy in PNs. Some factors, such as the LIM cofactor Chip, appear to affect wiring in all PN classes . Other TFs show lineage specific restriction in expression and regulatory effects. For example, the LIM-homeodomain TF Islet is required for proper targeting of a subset of adPNs and lPNs, while the homeodomain TF Cut is required in only a subset of lPNs and all vPNs . As another example, POU-domain TFs Acj6 and Drifter have restricted expression patterns in adPN and lPN lineages, respectively, and control wiring specificity in their respective lineages . A recently identified BTB-Zn-finger TF, Chinmo, regulates birth order-dependent wiring of adPNs. Loss of chinmo results in adPNs born in early larval life acquiring the targeting specificity of late-born PNs within the same lineage . There are also TFs that appear to affect targeting of a single PN type. For example, the LIM-homeodomain TF Lim1 is necessary for proper targeting of a single vPN to the DA1 glomerulus, and is regulated by Cut. The Zn-finger TF Squeeze appears to be necessary for the innervation of a single lPN glomerulus, DM5 . All these studies support a model where the targeting specificity of a particular PN is regulated by a unique complement of TFs, and additional members of the TF code remain to be identified.
The gene longitudinals lacking (lola) encodes a molecularly diverse BTB-Zn-finger TF with at least 20 unique protein isoforms (Figure 1b). Each isoform is formed by combining a set of common BTB-containing amino-terminal exons to unique Zn-finger-containing carboxy-terminal exons via trans- and/or cis-pre-mRNA splicing [16-18]. The BTB (Broad complex, Tramtrack, Bric à brac) domain, also referred to as the POZ (poxvirus and Zn-finger) domain, is a common domain likely involved in protein-protein interactions . Most lola isoforms have one or more unique Zn-fingers of either the C2H2 type that binds DNA, or the unusual C2HC class that binds nucleosomes, non-histone proteins, RNA and DNA [20,21]. At least some Lola isoforms bind DNA directly . Interestingly, three Lola isoforms lack Zn-fingers and theoretically could be involved in heteromeric regulatory interactions with other Lola isoforms, as Lola was found to bind itself in yeast-two-hybrid interactions and co-immunoprecipitation experiments [21,23]. Lola also likely interacts with other proteins such as chromosomal kinase JIL-1 .
Lola was initially identified as a factor that regulates axon guidance in the embryonic central nervous system longitudinal tracks, and is reported to function in a wide array of other cellular processes. lola mutants exhibit defects in the extension of embryonic longitudinal axons and midline crossing, orientation of lateral chordotonal neurons, and ISNb axon growth and elaboration [24,25]. Mutation of Lola isoforms K or L is sufficient to inactivate a specific subset of lola functions in ISNb neurons, suggesting that different Lola isoforms may have unique functions . lola was recently identified to disrupt ORN axonal innervation of the AL in an overexpression screen . lola may exert its effect through transcriptional regulation of cell-surface molecules, and has been reported to genetically interact with Notch, slit, and robo [27,28]. lola may have a more general regulatory role as a polycomb group (PcG) factor affecting cell proliferation via the Notch pathway and regulating wing development through a genetic interaction with cut .
In this study, we show that lola plays an essential role in PN identity and wiring specificity. lola appears to be a general factor that affects wiring of both axons and dendrites in all three lineages of PNs. Overexpression of UAS-lola T and UAS-lola L, but not UAS-lola A, results in wiring defects. Additionally, expression of single lola isoforms is insufficient to rescue the lola null phenotype and often causes additional defects specific to the isoforms expressed, suggesting the importance of Lola molecular diversity in regulating PN wiring specificity. Indeed most lola isoforms are expressed in PNs but at different levels. Finally, consistent with previous findings of transcriptional regulation and potential PcG function, we find that lola likely regulates lim1and several Gal4 enhancer trap lines in PNs. We suggest that lola regulates PN wiring through transcriptional regulation of downstream targets involved in defining neuronal identity and targeting specificity.