Wiring specificity in Drosophila olfactory projection neurons is regulated by intrinsic TFs [13-15]. Here we have characterized an additional TF, Lola, as necessary for proper AL wiring specificity of both PN dendrites and axons. lola-/- mutant clones show a wide variety of wiring phenotypes, including a loss of targeting to correct glomeruli, innervation of ectopic targets, a loss of lineage restriction in dendritic projections, a loss of AL boundary restriction, and an increase in wandering projections and ectopic branch formation. Additionally, lola appears normally to repress the expression of multiple Gal4 drivers in certain cell-types and positively regulate lim1 in vPNs. These data suggest that lola regulates multiple developmental processes, including identity as well as wiring specificity of PNs, consistent with its expression in neuroblasts and post-mitotic neurons. In fact, the wide variety of phenotypes observed in lola mutants may be a general characteristic of genes involved in multiple developmental processes, as similar phenotypes are observed in mutants of general TF complex cofactors, such as chip [14], or are known chromatin regulatory genes (D Berdnik and LL, unpublished data). This is the first report linking roles for lola in both fate and wiring specificity in the same cell-type, suggesting that PNs have a tightly linked relationship between targeting specificity and cell identity.
lola is a highly complex genetic locus and encodes at least 20 different splice isoforms [16]. Although Lola is expressed throughout development, quantitative RT-PCR reveals that individual isoforms have different levels of expression. The diversity of lola isoform expression in PNs also appears important, as expression of a single UAS-lola transgene fails to rescue null mutant phenotypes and often results in even more severe phenotypes that are specific to the expressed lola isoform. We tested available isoform-specific lola alleles, and found that an isoform L (but not K) mutant has mild targeting defects. The difficulties of generating other isoform-specific alleles in this locus, coupled with the strong gain-of-function effect of expressing transgenes of a single isoform, made it difficult for us to investigate further isoform-specific functions of lola in PN development. However, taken together with previous reports [16,25,16,25], our results support the notion that different isoforms may have unique functions in PNs in addition to embryonic motor neurons.
Several lines of evidence suggest that Lola can regulate chromatin structure. Loss of function (LOF) mutations in lola enhance the Pc/+ phenotype and result in a significant increase in the number of sex-combs on the second leg, as do other known PcG factors [28]. This reported PcG interaction, in addition to direct binding of the chromosomal kinase JIL-1 and the presence of the nucleosome binding C2HC class Zn-finger [20,21], suggests that at least some Lola isoforms may be part of chromatin regulatory complexes. Additionally, at least one Lola isoform has been shown to bind directly to DNA and regulate expression of the copia retrotransposon [22]. Many chromatin regulators are integral components in pathways of cell fate specification. Neuronal cell fate is closely linked with dendritic and axonal targeting, particularly in PNs where cell identities based on birth order are manifested as wiring choices and target specification [2]. This link between targeting and cell identity has been suggested in the MB, where mutation of the chromatin modifier polyhomeotic causes MB neurons to randomly express certain Hox genes that are normally suppressed and to display a wide variety of axon targeting defects that are neither a complete targeting fate switch nor consistent between individuals [32]. This is reminiscent of the lola phenotype in PNs, suggesting that lola might function in a similar manner.
lola may additionally play a more direct role in targeting. lola mutants affect some glomeruli more severely than others and have post-mitotic functions in DL1 dendritic targeting, suggesting that there is a differential requirement for, or sensitivity to, the loss of lola between individual PNs. lola also has specific targeting phenotypes in the embryonic central and peripheral nervous systems, as mutations in lola seem to disrupt axon guidance and extension of the Ap+, VUM, and ISNb neurons without affecting the numbers or differentiation of these neuronal subtypes [16,24,25,29]. Taken together, this evidence suggests that lola may participate both in fate determination and, more directly, in wiring specificity. If the processes of fate determination and wiring specificity are separable, one might expect a factor that controls wiring to directly regulate the expression of genes that function in targeting, such as cell surface receptors, while a factor that controls cell fate should affect the expression of 'fate' markers that subsequently disrupt downstream expression of the components of the wiring machinery in an indirect fashion. Future identification of direct targets of lola regulation will be informative and necessary to understand the exact mechanism of lola function in PNs.
The idea that TFs can differentially control fate specification and wiring specificity is consistent with the model of a hierarchical TF code that determines wiring specificity. It is interesting to consider the possibility that some TFs may be used at multiple points during identity and wiring specification. In neuron sub-type specification in the ventral nerve cord, the TF collier specifies the precursors of the Ap neuron subset, but must be turned-off in three of the four Ap cells as subsequent participation of collier in a multi-protein complex later in development defines a unique sub-type of Ap neurons [33]. During development of the peripheral nervous system, cut specifies sensory neuron identity and is required again later in development at different levels in subpopulations of dendritic arborization neurons to specify subclass arborization [34]. Similarly, in vPN development, the TF Cut is necessary in the neuroblast for generation of the correct number of vPNs and postmitotically for specific targeting of the VA1lm glomerulus [14]. Temporal regulation and multi-functionality of TFs are possible mechanisms to limit the required number of genes in a TF code while still uniquely specifying cell identity. Lola is a potential candidate for a TF that may function in this way, both specifying cell identity and playing more specific roles later in development in regulating wiring specificity.
Answers to all your Biology Questions
