A novel post-developmental role of the Hox genes underlies normal adult behavior

Significance At the end of development, cells must deactivate the genetic programs that guided their differentiation and switch on the molecular code script that ensures cellular stability and physiology. Within the nervous system, while there is a substantial understanding of how genes influence neuro-developmental processes, the genetic programs underlying cell function and stability in adult post-mitotic neurons remain largely unknown. Here we investigate this problem in Drosophila and discover that the Hox genes—which encode a family of evolutionarily-conserved developmental regulators, key for axial patterning—are essential for the normal physiology of post-mitotic neurons and adult behavior. Based on the evolutionary conservation of the Hox genes, we suggest that they may also play key neurophysiological roles in the adult forms of other species, including humans.

indicate that normal expression of Ubx in dopaminergic neurons is required for normal flight maintenance in adult flies.
The functional experiments described above implied that Ubx might be expressed in the adult dopaminergic system. Ubx is expressed in the adult ventral nerve cord (VNC) -the neural equivalent of the mammalian spinal cord (18) -across extensive domains on its dorsal and ventral aspects (Fig. 1A, and previous observations (3,10)). To probe specific expression within the dopaminergic system, we labelled the TH domain using nuclear GFP (NLS::GFP) under a dopaminergic driver (TH-Gal4) ( Fig. 2A), and mapped Ubx expression over this territory. The results of this experiment show that Ubx protein was indeed expressed in approximately ~ 30% (5±0.9) and 60% (17±0.5) of all adult dopaminergic neurons within the dorsal and ventral VNC regions, respectively (Fig. 2B). Furthermore, dopaminergic-specific conditional reduction of Ubx expression triggered at adult eclosion time (TH-Gal4; tub-Gal80 ts ; UAS-Ubx RNAi ) resulted in ~50% downregulation of Ubx expression within the TH domain after two weeks of treatment ( Fig. 2C-E). These expression analyses confirm that Ubx is normally expressed in adult TH neurons, and that our post-developmental manipulations of Ubx expression lead to a substantial reduction in protein formed within the dopaminergic system.
We then examined the relation between the activity of the dopaminergic system and flight control using optogenetics (19). For this, we firstly expressed CsChrimson (19) in the TH-domain and monitored effects on flight, observing that optogenetic activation of TH neurons (N.B. in the absence of an air-puff) is sufficient to initiate flight ( Fig. 3A and Movie 4). Secondly, we optogenetically inhibited TH neurons using GtACR (20) and observed that this treatment is sufficient to significantly reduce wing flapping ( To gain further insight on the properties of Ubx + dopaminergic neurons in connection to flight control, we developed a split-Gal4 approach (21). For this, we expressed complementary forms of the Gal4 transcriptional activator from two distinct promoters: Ubx (Gal4.DBD::Zip-) and pale (ple, the gene that encodes TH) (Gal4.AD::Zip + ) in order to reconstitute functional Gal4 protein only at the intersection between the Ubx and TH transcriptional domains (  (Fig. S4A,B), we sought to determine the natural patterns of activity of TH neurons in relation to normal flight, and how these were affected when Ubx expression was reduced, exploring the possibility that changes in Hox expression in the adult, might affect neural physiology. For this we made use of the CaLexA system (calcium-dependent nuclear import of LexA) (Fig. 4A) (23), which acts as a recorder device of neural activity in the tissue/cellular context of choice, and, thus, provides a platform to correlate neuronal activity patterns with specific behaviours (24). In a first series of experiments, we drove the CaLexA reporter from the TH-Gal4 driver and observed that after 30-min of constant flight there is a significant increase of CaLexA signal in the dopaminergic system when compared to a resting condition (Fig. 4B-E). Indeed, CaLexA signal intensity within the TH domain, is positively correlated with flight duration (Fig. 4F) indicating that the dopaminergic system is active during flight. In a second series of experiments, we reduced Ubx expression (by means of Ubx RNAi ) within TH neurons (although these experiments were conducted across the whole TH domain, expression of Ubx could have only been reduced in those cells that normally express the gene), and observed a significant decrease of neural activity in both, the second (T2) and third (T3) thoracic segments ( Fig. 4G-I). These data strongly support the notion that a reduction of Ubx expression in dopaminergic neurons leads to a decrease in neural activity within this domain. Interestingly, although we can indeed detect dopaminergic neurons in the T1 segment, as well as in the brain (Fig. S4C), we do not observe any CaLexA signal after flight in these regions. Further results indicate that flight-related dopaminergic activity takes place primarily within the VNC in segments T2-T3 (Fig. 4J), the thoracic regions where neural expression of Ubx is prominent (3,10). Expression of the CaLexA reporter within the intersectional domain THÇUbx reveals that even within this restricted aspect of the dopaminergic system, there are significant differences in CaLexA signal detected after flight, when compared with a resting control ( Fig. 4K-M). Furthermore, expression of the bacterial voltage-gated sodium channel NaChBac (25,26) within dopaminergic neurons expressing reduced levels of Ubx, leads to a rescue of the flight phenotype observed after TH-driven Ubx RNAi , as assessed by two independent behavioural tests: tethered flight ( Fig. 4N and Fig. S4D) and forced flight (Fig. 4O); this latter observation strongly suggests that a reduction of Ubx affects the physiological properties of the neuron, rather than its morphology. Altogether, the data above indicate that normal Ubx expression within the thoracic dopaminergic system is required for flight-related neural activity in thoracic TH neurons.
To probe the relation between the dopaminergic system and the flight motor system underlying normal flight, we first considered the (unlikely) possibility that thoracic TH neurons innervate flight muscle directly. Expression of membrane-bound GFP in all TH neurons shows that they do not produce any direct contacts with the flight motor (Fig. 5A) indicating that dopaminergic neurons must be communicating with the flight muscle system indirectly. To test this model, we To advance the understanding of the mechanisms that link a reduction in Ubx expression with the observed physiological change in TH neurons with impact on flight performance (Fig. 1, and Fig.   4), we conducted an RNA-sequencing experiment aimed at determining the transcriptome of TH neurons from the adult VNC. For this, we used FACS to isolate populations of TH neurons from the VNC expressing normal or downregulated Ubx (TH>Ubx RNAi ), extracted RNA, and compared the resulting transcriptomes using RNA-seq (Fig. 6A). Using edgeR analysis(32) we identify 233 differentially expressed genes (DEGs) (out of 5709 total genes detected) in TH>Ubx RNAi neurons relative to wt neurons (p-value <0.01; Table S1) (Fig. 6B). Using the DAVID platform (Database for Annotation, Visualization and Integrated Discovery(33)) we established the functional biological properties of the 233 DEGs. Looking for candidate genes that might perform Ubxdependent physiological roles in TH neurons, we noted that amongst the top-ten DEGs (Fold change (FC) >5) were four previously uncharacterised ion transport genes, with predicted symporter activity related to sodium, calcium, or potassium ion transport (Fig. 6C) (i.e. Gene Ontology (GO) terms enriched, Holm-Bonferroni test, p-value<0.05). These genes, which are substantially downregulated in response to a reduction of Ubx expression (≥10 FC, p<0.004; Fig   6D), include: CG1090, CG5687, CG9657 and CG6723 (Fig. 6D). Gene tree (Fig. 6E), protein alignment analysis ( Fig. S6 and S7), and protein structural predictions ( Fig. 6F and Fig. S6,7) using EMBL-EBI Clustal Omega, Jalview program(34), and JPred secondary structure prediction programs (35) for these symporter genes reveals that they belong to two independent lineages within the Solute Carrier gene (SLC) family: SCL5 and SCL24 (36-39). Although these genes have not been previously characterised in flies, the SLC24 genes have been shown to encode a diverse group of Na + /Ca 2+ -K + exchangers (NCKX)(37) in other species, and have been previously shown to play roles in nutrient sensing and sleep control in insects and mammals (36,37,39,40), and their dysfunction is correlated with neurological disease (36, 41, 42). Furthermore, expression data from the FlyAtlas expression database (43), reveals that these four symporter genes are primarily expressed in the Drosophila CNS, including the VNC (Table S2). Put together, these features suggest the hypothesis that SCL genes might play functional roles in the fly nervous system and might be some of the mediators through which Ubx exerts its roles on flight. To test whether normal expression of these symporter genes in TH neurons is required for flight maintenance, we examined flight behaviour (tethered flight and forced flight) following TH-specific RNAi-mediated downregulation of these genes. Our results show that TH-specific expression of RNAi constructs against CG1090 and CG6723 lead to an impairment of flight maintenance and ability ( We next, lifted the lid of the arena and after 2 min take-off was noted as successful for the flies that did not land within a radius of 7 cm. Optogenetic experiments were conducted by adapting Flypi device (53). For neuronal activation (CsChrimson, Pwr590) and inhibition (GtACR, Pwr470) a Neopixel twelve light-emitting diodes ring was positioned face-down around the infrared-capable camera objective about 3 cm above the tethered flies. Flies were recorded with lights off (in dark) and the response of the animal (males and females) were analysed during the following light on period. We exposed flies to approximately 4.9W/cm 2 for stimuli between 500-1000 ms using a custom-written Graphical User Interface (53). For all optogenetic activation experiments, adult flies upon eclosion were kept for 7-8 days before the experiment on food containing 0.5 mM all-trans retinal (Sigma).

Climbing assays
Startle-induced negative geotaxis (SING) test was used as climbing assay, and was monitored as Probes, 1:1000). Images were acquired with a Leica SP8 confocal microscope, processed, and analysed using FIJI ImageJ.

Calcium activity recording in thoracic muscles and CaLexA assays
Calcium activity recording in thoracic muscles was conducted on tethered fly, and calcium signalling recorded using epifluorescence microscopy Leica DM6000. Next, FIJI software, and subsequently Igor program were used to quantify calcium signal and generate traces, respectively.
The CaLexA (Calcium-dependent nuclear import of LexA) technique (23)  and Bioinformatic analysis using Drmelanogaster6 as reference for processing and alignment. Data was extracted as Total read count per gene, read per gene Length (Kp) -RPK, and RPK/million Mapped reads -RPKM. Genes with 0 reads and genes with more than 2-fold difference between biological duplicates were excluded. Lastly, fold-change between samples were calculated as a ratio of experimental samples over wild-type ones (experimental/wild-type).

Dendrogram generation and homology analysis
We used EMBL-EBI Clustal Omega (http://www.ebi.ac.uk/Tools/msa/clustalw2/) to conduct protein sequence alignment for with the symporter genes. We used the Jalview                   Ubx RNAi

A B
Issa & Alonso Figure S8