The emergence of appropriate behavior depends on the maturation of neuronal circuits. For this to occur the constituent neurons must not only develop a specific pattern of dendritic branching and synaptic connectivity, but must also express a stereotyped mix of ion channel genes that, together, set membrane excitability. Previous work in Drosophila has shown that motoneuron axon pathfinding is determined by a combinatorial code of transcription factor expression, including Eve, Islet and Hb9 [14–16, 18, 33]. In the present study, we show that at least one of these transcription factors, pivotal for morphological development, is also sufficient to affect the acquisition of electrical properties. It is well established that cellular morphology and electrical properties are key contributors to the diversity in neuronal signaling observed in the CNS. Against this backdrop it is gratifying that specific transcription factors are clearly able to regulate both these facets, which suggests that each are regulated by common developmental mechanisms.
Our electrophysiology clearly demonstrates that over-expression of eve in aCC/RP2 results in a down-regulation of both peak IKfast and amplitude of cholinergic minis. IKfast is a composite of at least two separate conductances encoded by shal and slo [19, 20, 34]. Addition of ChTX indicates that the effect of Eve on IKfast is through a reduction in the Slo-mediated component of this current. This conclusion is reinforced by two complementary approaches: DamID and QRT-PCR, the latter of which shows that slo is specifically repressed by Eve while shal remains unchanged. That both DamID and QRT-PCR identify slo as a target of Eve repression indicates that regulation of this ion channel gene is direct (that is, not occurring indirectly through unknown intermediates). Transcriptional regulation of slo is relatively complex , with at least five different promoters characterized: C0 and C1, which drive expression in neurons; C1b and C1c, which drive expression in the mid-gut; and C2, which drives expression in muscle and trachea. Each promoter gives rise to different splice variants. It remains to be determined how Eve regulates the expression of individual splice variants. The family of channels of which slo is a member, BK Ca2+-gated potassium channels, are activated by membrane depolarization during action potential firing and, as such, contribute to the after-hyperpolarization . Consistent with our data, Ikslo has been shown to regulate excitability via both duration and frequency of action potential firing  and also to modulate synaptic release .
In addition to decreasing peak IKfast, over-expression of eve in aCC/RP2 is sufficient to decrease the amplitude of action potential-dependent excitatory synaptic currents. Analysis of spontaneous minis reveals a similar reduction, which is indicative of reduced postsynaptic sensitivity to ACh. Our analyses indicate that expression of nAcRα-96Aa, a nAChR subunit, is negatively regulated by Eve. Unlike slo, however, we have not been able to use additional pharmacology to further corroborate this regulation. To show specificity, however, we used QRT-PCR to analyze mRNA levels for a second nAChR, nAcRβ-64B, a subunit not found in the same complex as nAcRα-96Aa . We saw no significant alterations in nAcRβ-64B mRNA levels, which allows us to tentatively conclude that Eve acts selectively on nAcRα-96Aa. This regulation, which is evident from DamID, QRT-PCR and electrophysiology, is again likely to be direct.
There is, however, a caveat that should be borne in mind in the interpretation of our data. This is that the changes we observed following over-expression of eve may be the average of both direct and indirect effects. We consider that indirect effects are likely to arise through two main sources. The first possibility would be changes that result from homeostatic compensatory mechanisms that are known to be active in Drosophila motoneurons [4, 22]. Such homeostatic mechanisms are capable of adjusting the relative peak amplitudes of specific membrane conductances (most notable INa and IK) to maintain consistency in action potential firing [4, 40]. Thus, changes to membrane excitability mediated by eve over-expression may be countered, at least in part, by homeostatic regulation. The effects observed in action potential firing following over-expression of eve in aCC/RP2 are in keeping with the known homeostatic mechanisms active in these two motoneurons; a decrease in exposure to synaptic excitation results in a compensatory increase in excitability to fire action potentials . The changes to the underlying electrical properties that are associated with this response (that is, a decrease in IKfast) are not, by comparison, in keeping with known homeostatic regulation that would predict significant increases in INa, IKfast and IKslow. Thus, whilst homeostatic regulation may contribute to the changes observed in membrane excitability, it seems unlikely that such mechanisms can account entirely for the changes observed in specific underlying membrane conductances.
A second possibility that may result in indirect change to membrane excitability is an alteration in the synaptic connectivity of motoneurons in which eve is over-expressed. Considering the documented role for eve in axon pathfinding, it is conceivable that alterations in electrical properties may arise as secondary consequences of morphological changes. We tend to discount this possibility for two reasons. First, all motoneurons were filled during our recordings and no obvious perturbation in wild-type morphology was noted. Of course, our level of analysis would not have detected subtle changes in fine dendritic branching. Second, Landgraf et al.  show that misexpression of eve is sufficient to only perturb axon pathfinding but not final target recognition in motoneurons (which is only delayed). Thus, although over-expression of eve in the ventral motoneurons is sufficient to misdirect their axons to dorsal muscles, these errors are rectified by the time of hatching.
In addition to our own experiments, there is a precedent for transcription factor misexpression resulting in altered membrane excitability. For example, expression of lox1, a homologue of the Drosophila genes sex-combs reduced and antennapedia, in leech neurons is sufficient to increase the size of the AP . In keeping with a basic principle of a combinatorial code, these authors report cell-type specific effects – only discrete groups of cells exhibit increased action potential amplitude. In the ascidian, Halocynthia roretizi, motoneurons that express the transcription factor HRlim have a set of common properties, including the expression of the sodium channel gene TuNa2 . Ectopic expression of HRlim is sufficient to drive expression of TuNa2 in cells that normally do not express either.
An important question raised by our examination of the consequences of eve over-expression in differing motoneuron subpopulations is the universality of its effects. Our data are consistent with a model in which the transcriptome of specific cell types is a critical determinant of the nature of regulation. Comparison of the effects produced by over-expression of eve in aCC/RP2 compared to those produced in the RPs shows differences consistent with this hypothesis. Whatever the precise mechanism, this apparent dichotomy of action strongly implicates a context-dependent mode of regulation. Such a conclusion is not too surprising given that Eve, Isl and Hb9 form part of a combinatorial code of motoneuron specification  and that specification of other neuronal properties, for example, dFMRFa expression and furnin 1 expression, is also under complex combinatorial and non-combinatorial control . Thus, the electrical properties of neuronal membranes are most likely dictated by a combined activity of a range of transcription factors, which our data indicates includes Eve, Islet and HB9. Alteration of this balance within specific neurons will likely result in an equally diverse range of effects on both electrical and signaling properties.